That is the argument at the centre of this episode — and in seven words, it is the entire thesis. Everything else is evidence.

Last week, in Part 1, I traced the line from the Strait of Hormuz to the server rack and argued that the post-war bargain between Europe and America has inverted. I used the analogy of a house — joists, plumbing, planning permission, landlord — to map the four chokepoints that converged in a single fortnight.

This week, I go underneath the house. To the foundations. Because the entire European technology sovereignty project — the Chips Act, the data sovereignty regulations, the cloud procurement frameworks — is a house being built on foundations nobody has paid for.

The bill has just arrived.

What this episode covers

This is a 31-minute solo episode. I work through four arguments, each building on the last:

The physics of the problem. Why an aluminium pot line cannot be cold-started — the cells are destroyed, not paused, when the cryolite solidifies at 960 degrees Celsius. Why every helium atom in commercial use was mined, not manufactured. Why the loss of three million tonnes of Gulf aluminium capacity and 25 per cent of global helium supply is a sovereignty story, not a commodity story.

The inconvenient truth. How Ricardo’s comparative advantage theory — honest, functional, and correct for thirty years — rested on three assumptions that this fortnight removed simultaneously. Why Adam Tooze calls it “zombie globalisation.” Why Matt Stoller’s financialisation critique explains how we got here.

The policy gap. The EU Chips Act’s three structural gaps: no allocation for primary aluminium processing, no strategic helium reserve (despite helium being on the Commission’s critical raw materials list since 2020), and no crisis-response mechanism for disruptions originating upstream of the fab. The UK Critical Minerals Strategy’s order-of-magnitude funding shortfall — £50 million committed against a problem that costs £8 to £15 billion.

The honest answer. A five-point metal floor strategy for the European bloc, extending Ed Conway’s Material World thesis to argue that digital sovereignty is a subset of material sovereignty. Total cost: between 40 and 60 billion euros over a decade. The number nobody has published yet.

The episode closes with a prediction — within 18 months, a major European fab will publicly disclose a helium allocation constraint delaying capacity expansion — and three amendments to the quarterly board paper introduced in Part 1.

Read the full written analysis

The written article — with all 17 endnotes, the image brief, and the predictive judgement with explicit falsifiability conditions — is here:

If you missed Part 1:

### Sources cited in this episode

[1]: Rio Tinto Alcan. “Sale of Lynemouth smelter to Klesch Group completed.” 2012 corporate archive, and “Rio Tinto sells Alcan Aluminium UK to Liberty House.” Financial Times, 16 December 2016.

[2]: Reuters. *“Gulf aluminium output crippled as Iranian strikes hit EGA and Alba facilities.”* 11 April 2026. See also CRU Group, *Aluminium Market Monitor*, April 2026 update.

[3]: Emirates Global Aluminium plc. *“Annual Report 2025 — European Customer Disclosures.”* https://www.ega.ae/en/investor-relations; International Aluminium Institute, Primary Aluminium Production Statistics Q1 2026

[4]: U.S. Geological Survey. “Mineral Commodity Summaries 2026 — Helium.” January 2026. ; Linde plc, Helium Supply Chain Disclosures, 2025 Annual Report; Air Liquide SA, Critical Gases Supply Position 2025.

[5]: David Ricardo. *On the Principles of Political Economy and Taxation*. John Murray, 1817. Chapter 7, “On Foreign Trade.”

[6]: Adam Tooze. *Chartbook* (Substack). See in particular the ongoing analysis of “zombie globalisation” across Chartbook issues 2023–2026.

Chartbook

A newsletter on economics, geopolitics and history from Adam Tooze. More substantial than the twitter feed. More freewheeling than what you might read from me in FT, Foreign Policy, New Statesman.

By Adam Tooze

[7]: Matt Stoller. *BIG* (Substack). See ongoing analysis at

BIG by Matt Stoller

The history and politics of monopoly power.

[8]: European Union. Regulation (EU) 2023/1781 of the European Parliament and of the Council of 13 September 2023 establishing a framework of measures for strengthening Europe’s semiconductor ecosystem (Chips Act).[9]: European Commission. “Chips Act Pillar II Implementation Report — First Annual Review.” March 2026. See Commission Staff Working Document SWD(2026) 047 final.

[10]: European Commission. “Critical Raw Materials Act — List of Critical and Strategic Raw Materials.” Regulation (EU) 2024/1252, Annex I. Helium listed as critical since 2020 Communication COM(2020) 474.

[11]: Estimate synthesised from European Round Table for Industry (ERT) position papers on industrial policy 2025, Bruegel working papers on European strategic autonomy, and off-the-record conversations with Commission officials. Not yet formally published by any Commission document.

[12]: HM Government, Department for Business, Energy and Industrial Strategy. *“UK Critical Minerals Strategy.”* July 2022; refreshed March 2023 as “Critical Minerals Refresh.”

[13]: Tata Steel UK. *“Port Talbot Transformation Announcement.”* January 2024.

[14]: HM Government, Department for Business and Trade. “Invest 2035: The UK’s Modern Industrial Strategy.”* Green paper, October 2024.

[15]: Estimate synthesised from CRU Group cost models for greenfield primary aluminium facilities, 2025 capital cost indices, and comparable recent projects (Rio Tinto Kitimat modernisation, Emirates Global Aluminium Al Taweelah expansion).

[16]: Ed Conway. *Material World: A Substantial Story of Our Past and Future*. W.H. Allen, 2023. See also ongoing analysis at Material World Substack:

Listen and subscribe

YouTube:

Apple Podcasts:

Spotify:

Substack: https://thecontrollayer.arkava.ai

Subscribe on the Substack to get every episode and written analysis by email — free.

Join the conversation

The metal floor argument is the one I expect to generate the most disagreement in this series. The numbers are large. The policy gap is structural. And the question of whether Europe should rebuild domestic processing capacity or find a different kind of resilience is genuinely open.

I want to hear from people working in industrial policy, commodity trading, semiconductor procurement, critical minerals strategy, or defence supply chains. Where am I wrong? Where is the five-point strategy naive? What am I missing?

Drop a comment below, or find me on LinkedIn

Amer Altaf is Founder and CEO of Arkava and Managing Editor of The Control Layer.

Aluminium and helium are the supply chain layer that cannot be digitised. Neither the EU Chips Act nor UK industrial policy has yet costed an honest answer. Part two of the Four Chokepoints series.

For thirty years, the answer to “why is aluminium smelted in the Gulf rather than in Northumberland?” was the same answer we gave for every offshored process: cheaper labour, cheaper energy, cheaper environmental compliance. The answer was honest, until this April.

The reason the aluminium is not smelted in Northumberland is not that it cannot be — the British Aluminium Company ran Lynemouth until 2012 and Fort William until Rio Tinto sold it to Liberty House in 2016, and the aluminium chemistry does not care which country’s grid it draws from.[1] The reason is that for three decades the financial logic of global trade treated processing location as an arbitrage opportunity, and Western policy treated resilience as a free good. When labour cost gaps of 70 per cent and energy cost gaps of 40 per cent are set against a supply chain risk premium of zero, the location decision makes itself.

This April, the risk premium moved off zero. What did not move — what cannot move — is the location of the ore, the chemistry of electrolysis, or the geology of helium. The metal floor, the supply chain layer on which every software sovereignty intervention Europe is currently proposing quietly rests, has always had a physical precondition. It is only this fortnight that the precondition has become legible on a quarterly risk paper.

This is what European industrial policy has not yet costed.

Share The Control Layer

What actually broke

The aluminium disruption is precise. Iranian strikes on Gulf smelting infrastructure removed approximately three million tonnes of annual primary aluminium capacity from Emirates Global Aluminium at Jebel Ali and Aluminium Bahrain at Sitra, with both facilities now operating under partial force majeure.[2] Restart timelines sit between six and twelve months, and the reason is chemical rather than structural. An aluminium pot line is a row of electrolytic cells running molten cryolite at approximately 960°C, drawing direct current at industrial scale continuously for the working life of the cell. When power or feed is interrupted for more than a few hours, the cryolite solidifies and the cell is destroyed. Restart requires rebuilding the cell, not restoring power to it. EGA and Alba together supply material into European automotive, aerospace, and construction grades — specialised alloys that are not fungible with Chinese export aluminium.[3]

The helium disruption is different in structure and in timescale. Qatar is the world’s second-largest helium producer, contributing approximately 25 per cent of global supply through the Ras Laffan complex, where helium is extracted as a by-product of liquefied natural gas production and shipped on the same LNG tanker fleet that now exits the Gulf through the contested waters of Hormuz.[4] US Geological Survey data places the remaining global supply at approximately 45 per cent from the United States, 5 per cent from Russia, and the balance from Algeria, Poland, Canada, and Tanzania.[4] Helium cannot be synthesised at industrial scale. Every helium atom currently in commercial use was mined. When the Qatari supply stalls, the semiconductor fabrication, magnetic resonance imaging, and cryogenic cooling sectors do not receive higher prices — they receive allocation letters. Strategic reserves at major fabs are measured in weeks.

These are not commodity stories. Commodity stories price in and price out. These are sovereignty stories, because the resolution of a sovereignty story is not a price — it is a policy.

The inconvenient truth global trade has underwritten

The economist in the room will point out that the labour arbitrage logic that moved aluminium processing to the Gulf, steel processing to China, and rare earth refining to Inner Mongolia was not a mistake. For thirty years, it was the orthodox application of comparative advantage theory — the argument, traceable to David Ricardo in 1817, that nations should specialise in what they produce relatively more efficiently and trade for everything else. The orthodoxy, applied to processing, said that if Bahrain could smelt aluminium at 60 per cent of the cost of Northumberland because its energy was subsidised by natural gas extraction that would otherwise be flared, the efficient outcome was to smelt the aluminium in Bahrain and ship the ingots to Europe.[5]

The orthodoxy was honest. It had an inconvenient truth buried in the first paragraph of every economics textbook that cited it: comparative advantage assumes frictionless trade between peaceful trading partners in a stable geopolitical order. Remove any of those three assumptions and the arbitrage stops being efficient and starts being dangerous.

Adam Tooze has described this as “zombie globalisation” — the post-1990 trading order that did not end in 2016 with Brexit, or in 2018 with the first Trump tariffs, or in 2022 with the invasion of Ukraine, but has been dying slowly since.[6] The aluminium and helium chokepoints exposed in April are one more organ failure in a patient the attending physicians have been reluctant to call.

Matt Stoller has argued, from a different direction, that the financialisation of Western industrial policy confused shareholder returns with national capability, and the consequence was the systematic under-investment in domestic processing infrastructure that we are now trying to price back in.[7] The argument here is not that the Ricardian orthodoxy was wrong in 1817. It is that the specific form of the orthodoxy applied between roughly 1995 and 2022 treated one assumption — stable geopolitical order — as permanent when it was contingent, and the entire architecture of Western processing location was built on that assumption.

The disruption of April 2026 is the moment the assumption stops being quietly permissible and starts being expensively inoperable. The question for UK and European policy is no longer whether to rebuild domestic processing capability. It is how much of the last thirty years’ worth of forgone investment has to be reabsorbed, and across what timescale.

What the EU Chips Act has actually costed

The European Chips Act, adopted in July 2023, committed €43 billion of public and private investment to European semiconductor sovereignty across three pillars.[8] Pillar I funds the Chips for Europe Initiative — pilot lines, design infrastructure, and skills — at approximately €3.3 billion from the Horizon Europe and Digital Europe programmes. Pillar II establishes a framework for state aid to first-of-a-kind facilities, which is the instrument under which the Intel Magdeburg and TSMC Dresden subsidies were structured. Pillar III coordinates crisis response.

The Chips Act is, on its own terms, a well-designed instrument. It has moved fabrication capability towards Europe. But its structural assumption — the one that aligns it with the last thirty years of industrial orthodoxy rather than the next thirty — is that the upstream material inputs to European semiconductor fabrication will continue to be sourced from global markets on something like the pre-2022 terms. The Act’s implementing documents treat bulk materials as a procurement problem for the fabricators rather than a sovereignty problem for the bloc.[9]

The gap is visible in three specific places. First, there is no Pillar II allocation for primary aluminium, specialty steels, or non-ferrous processing that feeds the fabrication estate. Second, there is no strategic helium reserve funded under the Act, despite helium’s classification as a critical raw material by the European Commission since 2020.[10] Third, the Act’s crisis-response mechanisms under Pillar III do not contemplate the scenario in which the disruption to semiconductor production originates upstream of the fab in the material supply chain, rather than at the fab itself.

These are not criticisms. They are gaps. The Act was built to solve the problem Europe had diagnosed in 2021 — we do not have enough leading-edge fabrication capacity — not the problem Europe has acquired in 2026 — we do not have enough upstream material sovereignty to keep the fabrication capacity running. The next Chips Act iteration, which Commission officials have begun briefing as Chips Act 2.0 in private settings, will need to add a fourth pillar addressing the metal floor. The current budget envelope will not stretch to it. The estimate Sarah’s board paper will want to see — which this piece cannot yet quote because no Commission document has published it — is somewhere in the region of €20 to €30 billion of new primary processing investment, phased over seven to ten years, to close the gap the Act as drafted left open.[11]

What the UK Critical Minerals Strategy has actually costed

The United Kingdom’s position is structurally weaker and financially smaller. The UK Critical Minerals Strategy, first published in July 2022 and refreshed in March 2023, identified 18 minerals as critical to UK supply chains and committed approximately £50 million to the Critical Minerals Intelligence Centre and allied research infrastructure.[12] Aluminium is on the list. Helium is on the list. The funding envelope is between one and two orders of magnitude smaller than what the underlying problem requires.

The UK’s processing capacity position is the direct inheritance of the deindustrialisation decisions of the 1980s and 1990s. Primary aluminium smelting in the United Kingdom effectively ended when Rio Tinto sold Lynemouth in 2012 and Liberty House took over Fort William in 2016 — the latter remaining operational but at a fraction of historical capacity.[1] Primary steel processing has contracted similarly, with the closure of the Port Talbot blast furnaces in 2024 accelerating the trend.[13] There is no domestic helium production in the United Kingdom at any commercial scale.

The Labour Government’s Invest 2035 industrial strategy, published in green paper in October 2024 and scheduled for white paper publication during 2025, identified advanced manufacturing and critical minerals as two of its eight growth-driving sectors.[14] The direction is correct. The funding is not yet specified. If the UK is to reopen primary aluminium smelting at strategic scale — not for commodity markets, but for semiconductor-grade and aerospace-grade allocation sovereignty — the capital cost is between £1.5 and £3 billion for a single modern facility, before the grid connection and the downstream alloy mill are priced.[15] The Invest 2035 envelope is not currently sized to absorb this.

The honest answer to “what would it cost the United Kingdom to stand up a sovereign metal floor?” is somewhere between £8 and £15 billion over a decade, across aluminium, specialty steels, helium storage, and the downstream processing that feeds the defence and semiconductor supply chains. That number is not in any published UK document. It is the number a senior civil servant in the Department for Business and Trade would give in private, off the record, if you asked them directly.

What an honest metal floor strategy would cost

Ed Conway’s Material World argues that the twentieth century was underwritten by six physical materials — sand, salt, iron, copper, oil, and lithium — and that the twenty-first century is underwritten by the same six plus helium, rare earths, and silicon.[16] The argument that this piece extends is that digital sovereignty, which is commonly treated as a software and cloud problem, is actually a subset of material sovereignty — and that the material substrate is the layer Western policy has most systematically under-priced.

An honest metal floor strategy for the European bloc would do five things, in approximately this order. It would establish strategic reserves for aluminium ingots, helium, and specialty steels at quantities measured in weeks rather than days of consumption. It would fund the reopening of primary processing capacity in three geographies — the United Kingdom at the former Fort William and Lynemouth sites, the northern European low-carbon energy corridor around Iceland, Norway, and northern Scotland, and the southern European renewable corridor around Iberia. It would establish long-dated offtake contracts between sovereign processors and Tier 1 European manufacturers — Airbus, Stellantis, Siemens, ASML — at prices that absorb the strategic premium explicitly rather than demanding that the processors absorb it invisibly. It would create a sovereign premium in public procurement, analogous to the “Buy British” or “Made in the EU” procurement preference, that makes the metal floor commercially viable without relying on tariff walls. And it would, finally, cost and publish the number honestly — which no current document does.

The total capital requirement for the European bloc and the United Kingdom combined is between €40 and €60 billion over a decade, on top of the existing Chips Act envelope.[11] That is the honest number. It is a politically uncomfortable number. It is also small relative to the €1.8 trillion NextGenerationEU budget and trivial relative to the aggregate

capitalisation of the European automotive and aerospace sectors that currently depend on the imported metal that the number would replace.

The helium prediction

Within 18 months — by October 2027 — a major European semiconductor fabricator will publicly disclose a helium allocation constraint that delays a planned capacity expansion by at least six months. The fabs to watch are Infineon Dresden, STMicroelectronics Crolles, Bosch Reutlingen, GlobalFoundries Dresden, and the downstream fab customers of ASML Veldhoven. The disclosure will appear in a quarterly earnings call, a capacity announcement, or a regulatory filing under European Securities and Markets Authority rules on material adverse changes.

When it appears, it will be reported as a supply chain story. Sarah’s board paper, if Part 2 has done its job, will recognise it as something else: the first legible instance, on a European earnings call, of the metal floor asserting itself. The capital re-allocation towards sovereign processing that follows — if the 18-month prediction from Part 1 also lands — will be the visible market response to a physical precondition that had been quietly absent from the risk model.

The bottom line

Three specific amendments to Sarah’s quarterly risk paper, based on what this piece has argued.

The vendor concentration matrix introduced in Part 1 needs a second tier: the upstream material concentration underneath each vendor. Any entry that reads “ASML” or “Intel Magdeburg” or “Infineon Dresden” must now carry a second line naming the material supply chains those vendors depend on, and the geographic concentration of those chains. Most current matrices do not have this second tier at all.

The capital expenditure forecast for the next five years needs a sovereign premium line. For any procurement category touching critical infrastructure, defence, or public services, a 15 to 25 per cent sovereign premium should now be modelled explicitly rather than treated as an anomaly. Sarah’s CFO will resist this. The resistance is the inherited reflex of thirty years of cost-optimisation orthodoxy, and it is now the expensive reflex.

The political risk register needs one new line item: European industrial policy coherence. Not geopolitical instability, which is vague. Coherence: the question of whether the EU Chips Act 2.0, the UK Invest 2035 white paper, and the various national critical minerals strategies converge on an integrated metal floor programme, or whether they remain four or five parallel under-funded efforts that each solve a different quarter of the problem.

The work of this series is to argue that these amendments are not radical — they are overdue. Part 3 will examine the equipment chokehold: ASML, the MATCH Act, and the end of the allied exemption that structured European semiconductor policy since the 1990s. The argument there will be that the metal floor is the precondition Washington cannot control, and the equipment chokehold is the precondition Washington is now attempting to weaponise.

The Control Layer exists to translate these conditions into the language of the board paper, in a voice a board will quote. If that is the translation you have been looking for, subscribe — and forward this piece to the colleague who has been asking the question you could not quite answer.

Next: Part 3 — The equipment chokehold: ASML, the MATCH Act, and the end of the allied exemption.

Subscribe now

References

[1]: Rio Tinto Alcan. “Sale of Lynemouth smelter to Klesch Group completed.” 2012 corporate archive, and “Rio Tinto sells Alcan Aluminium UK to Liberty House.” Financial Times, 16 December 2016.

[2]: Reuters. *“Gulf aluminium output crippled as Iranian strikes hit EGA and Alba facilities.”* 11 April 2026. See also CRU Group, *Aluminium Market Monitor*, April 2026 update.

[3]: Emirates Global Aluminium plc. *“Annual Report 2025 — European Customer Disclosures.”* https://www.ega.ae/en/investor-relations; International Aluminium Institute, Primary Aluminium Production Statistics Q1 2026

[4]: U.S. Geological Survey. “Mineral Commodity Summaries 2026 — Helium.” January 2026. ; Linde plc, Helium Supply Chain Disclosures, 2025 Annual Report; Air Liquide SA, Critical Gases Supply Position 2025.

[5]: David Ricardo. *On the Principles of Political Economy and Taxation*. John Murray, 1817. Chapter 7, “On Foreign Trade.”

[6]: Adam Tooze. *Chartbook* (Substack). See in particular the ongoing analysis of “zombie globalisation” across Chartbook issues 2023–2026.

Chartbook

A newsletter on economics, geopolitics and history from Adam Tooze. More substantial than the twitter feed. More freewheeling than what you might read from me in FT, Foreign Policy, New Statesman.

By Adam Tooze

[7]: Matt Stoller. *BIG* (Substack). See ongoing analysis at

BIG by Matt Stoller

The history and politics of monopoly power.

[8]: European Union. Regulation (EU) 2023/1781 of the European Parliament and of the Council of 13 September 2023 establishing a framework of measures for strengthening Europe’s semiconductor ecosystem (Chips Act).[9]: European Commission. “Chips Act Pillar II Implementation Report — First Annual Review.” March 2026. See Commission Staff Working Document SWD(2026) 047 final.

[10]: European Commission. “Critical Raw Materials Act — List of Critical and Strategic Raw Materials.” Regulation (EU) 2024/1252, Annex I. Helium listed as critical since 2020 Communication COM(2020) 474.

[11]: Estimate synthesised from European Round Table for Industry (ERT) position papers on industrial policy 2025, Bruegel working papers on European strategic autonomy, and off-the-record conversations with Commission officials. Not yet formally published by any Commission document.

[12]: HM Government, Department for Business, Energy and Industrial Strategy. *“UK Critical Minerals Strategy.”* July 2022; refreshed March 2023 as “Critical Minerals Refresh.”

[13]: Tata Steel UK. *“Port Talbot Transformation Announcement.”* January 2024.

[14]: HM Government, Department for Business and Trade. “Invest 2035: The UK’s Modern Industrial Strategy.”* Green paper, October 2024.

[15]: Estimate synthesised from CRU Group cost models for greenfield primary aluminium facilities, 2025 capital cost indices, and comparable recent projects (Rio Tinto Kitimat modernisation, Emirates Global Aluminium Al Taweelah expansion).

[16]: Ed Conway. *Material World: A Substantial Story of Our Past and Future*. W.H. Allen, 2023. See also ongoing analysis at Material World Substack:

Material World

Unexpected stories from the underbelly of the modern world

By Ed Conway

Author

Amer Altaf is Founder and CEO of Arkava, a UK and European sovereign AI agentic automation business, and Managing Editor of The Control Layer, the publication where he tracks the convergence of cyber security, technology sovereignty, and geopolitics. A techUK member, he contributes to industry engagement on UK technology sovereignty policy. His January 2026 analysis of Airbus’s migration off Google and Microsoft argued European cloud sovereignty had moved from policy aspiration to active corporate posture — six months before the Hormuz blockade made the argument unavoidable. He is currently writing on cloud security for Oxford University Press’s Expert Essentials series.

Article series

Series: Four Chokepoints, Part 2 of 5.

Previous instalment: Part 1 — Four chokepoints: inside the fortnight that made European technology sovereignty unavoidable.

Next instalment: Part 3 — The equipment chokehold: ASML, the MATCH Act, and the end of the allied exemption.

The headlines called it an oil shock. It isn’t. In this episode, I trace the line from the Strait of Hormuz to the server rack in your data centre — through the aluminium smelters, the helium tankers, the Dutch lithography plant, and the Franco-British communiqué that said something no European government has said out loud since 1945.

My argument: the post-war bargain between Europe and America has not strained. It has inverted. And the board papers being written this month still treat it as an oil shock.

If you prefer to read the full written analysis

— with all twelve endnotes and the complete image brief

What this episode covers

The four failures — what actually happened in the fortnight to 12 April 2026: three million tonnes of Gulf aluminium capacity offline, helium tanker disruptions threatening semiconductor fabs, the MATCH Act’s 150-day ultimatum to ASML, and the first Franco-British military operation outside the American framework since the post-war settlement.

The inversion — why Helen Thompson’s thesis about the contingent Atlantic settlement is no longer history but operational reality, and why Adam Tooze’s structural polycrisis framework is the correct lens for the events of this fortnight.

Three lines for the board paper — the vendor concentration matrix, the supply chain map, and the political risk register: what’s wrong with each and what they should say instead.

The eighteen-month prediction — procurement-driven re-sovereigntisation in UK and EU public-sector contracts, with explicit falsifiability conditions I’ll track quarterly on The Control Layer.

The Four Chokepoints series

This is Part 1 of a five-part series tracing the supply chain from the Strait of Hormuz to your procurement decision.

Part 1 — The Inversion (this episode): the fortnight, the four failures, and the board paper that’s wrong.

Part 2 — The Metal Floor: aluminium, helium, and the hidden physical inputs to the sovereign technology stack. Why the metal shortage is a financial choice, not a geological problem.

Parts 3–5 — forthcoming. Subscribe to receive each instalment.

Subscribe now

Listen and subscribe

- YouTube:

- Apple Podcasts:

- Spotify:

- Substack:

I want to hear your take. What did I get wrong? What did I miss? If you sit on an Audit and Risk Committee, does your board paper currently describe this as an oil shock — and if so, what would it take to change that?

Leave a comment below or find me on LinkedIn

The Control Layer is hosted by Amer Altaf, Founder and CEO of Arkava — Trusted Intelligence, Tangible Impact.

On the morning of 8 April, somewhere between the Strait of Hormuz and a Dutch lithography plant in Veldhoven, four supply chains began to fail at once.

The first was aluminium. Iranian strikes on Gulf smelting infrastructure removed approximately three million tonnes of annual capacity from Emirates Global Aluminium and Aluminium Bahrain, both now operating under partial force majeure with restart timelines of six to twelve months.[1] The second was helium. A significant share of the world’s helium supply exits Qatar’s Ras Laffan complex on the same liquefied natural gas tankers that transit Hormuz — the tankers that are now variously diverted, delayed, or waiting.[2] The third was semiconductor manufacturing equipment. A bill introduced in Washington the same week, the Multilateral Alignment of Technology Controls on Hardware Act, placed allied vendors including ASML on a 150-day clock to match US export controls or lose access to US intellectual property and components.[3] The fourth was the alliance itself. France and the United Kingdom announced a joint naval mission to the Strait of Hormuz operating outside the American blockade framework, the first operational admission that European strategic interests can no longer be safely delegated to Washington.[4]

Four chokepoints. One fortnight. Board papers still calling it an oil shock.

Share

What failed at the metal floor

Aluminium is the first chokepoint because aluminium is the first input. Data centre chassis, server rack framing, cable trays, electric vehicle battery housings, and the structural components of satellite and aerospace platforms all rely on specialised grades from Gulf smelters. The loss of three million tonnes of annual capacity from Emirates Global Aluminium and Aluminium Bahrain transmits into European automotive, aerospace, and construction sectors within a single inventory cycle.[1] Procurement directors who treated this material as commoditised will discover, over the next quarter, that it was a single-point dependency.

Helium is the chokepoint that almost nobody outside semiconductor and cryogenic engineering is currently watching. Qatar is the world’s second-largest helium producer — approximately 25 per cent of global supply, exiting via Ras Laffan on the same tanker fleet now caught in the Hormuz disruption.[2] Helium is not substitutable for semiconductor fabrication, magnetic resonance imaging, or the cryogenic cooling of superconducting infrastructure. A protracted closure does not simply raise prices; it triggers allocation rationing. Strategic reserves at fabs in Taiwan, South Korea, and the Netherlands are measured in weeks, not months.

The equipment chokehold

The Multilateral Alignment of Technology Controls on Hardware Act is a Foreign Direct Product Rule delivered as legislation. It gives allied vendors 150 days to match the US export control posture on semiconductor manufacturing equipment shipped to China, or lose access to US components, software, and intellectual property.[^3] ASML, the Dutch monopolist in lithography, derives approximately 50 per cent of its Deep Ultraviolet sales from China.[^5] The MATCH Act is not a trade dispute. It is the termination of the allied exemption that has structured European technology policy since the 1990s.

The administrative bottleneck running parallel to the legislation is the deeper signal. The US Bureau of Industry and Security, hollowed out by a 19 per cent reduction in headcount, now processes export licences for trusted allies — the United Kingdom, Japan, Canada — at an average of 76 days, double the 2023 baseline.[6] Multi-billion dollar backlogs in the deployment of artificial intelligence infrastructure are accumulating in countries Washington formally describes as partners. The semiconductor industry has formally warned that this undermines Western competitiveness; the political response so far has been silence.

When the alliance itself becomes a chokepoint

The Franco-British maritime mission is the piece of news that looked smallest and mattered most. By announcing a joint naval deployment to the Strait of Hormuz operating outside the American blockade framework, Paris and London formalised what had until then been a private frustration: European strategic interests can no longer be safely delegated to Washington.[4] The mission is described as “strictly defensive” — diplomatic language for we are not doing what the Americans are doing. The communiqué is the document historians will reach for when they date the inversion of the post-war bargain.

Spain’s concurrent state visit to Beijing — Prime Minister Pedro Sánchez meeting President Xi Jinping to secure Chinese foreign direct investment in Spanish renewable energy, biopharmaceuticals, and engineering firms — confirms that the fragmentation runs in multiple directions at once.[7] The European Union’s stated “de-risking” strategy is contradicted from inside the bloc the same week it is contradicted from outside. The €359.8 billion EU trade deficit with China published by Eurostat on 10 April is the empirical evidence that de-risking has not, in fact, been happening.[8]

The inversion Helen Thompson predicted

The historian Helen Thompson argued in Disorder: Hard Times in the 21st Century that the post-war Atlantic settlement was always a contingent arrangement of energy, security, and finance — not a permanent moral architecture, and not stable indefinitely.[9] What Thompson described in long historical register, this fortnight delivered in operational form. The bargain that afforded eighty years of European prosperity has not strained; it has inverted. The arrangement that once underwrote European security now actively imposes the principal structural risk to European prosperity, because the unilateral instruments Washington is willing to deploy — naval blockades, extraterritorial export controls, tariff revisions targeting allied vendors — operate on Europe with the same force they operate on adversaries.

The implication for any UK or European business is not that the United States has become hostile. It is that the United States has become unpredictable in ways that make American technology, American security guarantees, and American capital markets newly risky as default assumptions. A treasury function that holds a majority of its corporate cash in dollar instruments, a procurement function that defaults to American hyperscalers, an operations function that assumes American-flagged vessels will move its goods — each of these is now a position with a beta to American political volatility, which is currently the highest beta in the global system.

What your board paper should now say

Adam Tooze has argued, persuasively, that the polycrisis is structural rather than episodic — that the convergent shocks of the past five years are not a sequence of unrelated events but the surface expression of a single underlying re-ordering.[10] If that frame is correct, and the evidence of this fortnight is that it is, then the quarterly enterprise risk paper that treats geopolitics as a “watch item” under “external environment” is no longer a serious document.

Three lines that should be rewritten in the next paper to the Audit and Risk Committee:

The first line is the vendor concentration matrix. Any matrix that aggregates “cloud” or “AI infrastructure” as a single category, without breaking out the proportion held with US hyperscalers versus UK or EU sovereign providers, is now under-specified. The MATCH Act timeline is 150 days. The board needs to know, before that clock expires, which workloads are exposed to a unilateral American export control on the equipment used to build the infrastructure those workloads run on.

The second line is the supply chain map. Any map that treats freight insurance, lead times, and chokepoint exposure as logistics rather than as strategic risk is mis-categorising the data. The 7 to 15 days now added to Asia–Europe ocean transit via the Cape of Good Hope, the 10 per cent effective capacity reduction across global ocean networks, and the 30 to 40 per cent increase in greenhouse gas emissions on diverted routes are not operational frictions.[11] They are the new permanent baseline for any inventory model.

The third line is the political risk register. Any register that lists “geopolitical instability” as a single item, without naming the specific transatlantic instruments that now operate against allied interests, is performing risk theatre. The instruments are named: the MATCH Act, the Section 232 tariff revisions, the BIS licence delays, the Strait of Hormuz blockade. They are dated. They are quantifiable.

What happens in the next eighteen months

Noah Smith has argued that the West’s stance towards China is “derisking, not decoupling” — a graduated reduction of dependency rather than a clean break.[12] That framing was accurate for the first half of the decade. It is now, in the spring of 2026, behind the curve. We are past derisking. The next eighteen months are about procurement-driven re-sovereigntisation, and the evidence will accumulate in three places.

The first place is public-sector contract awards. UK Government Digital Service procurement, the Crown Commercial Service framework refreshes, the EU Chips Act Pillar II disbursements, and the German Bundesnetzagentur tenders for sovereign data infrastructure will publish award data quarterly. Watch the share going to UK and EU domiciled providers. The trend is already visible; the question is the slope.

The second place is capital re-allocation. Institutional investors and pension funds with ESG and political risk mandates will rotate from US tech megacaps into UK and European technology equities at the rate the procurement signal becomes legible. London is already the deeper liquidity venue for European listings; the question is whether sovereign procurement creates enough domestic earnings growth to make London the primary listing venue rather than the secondary one.

The third place is the regulatory perimeter. Labour has publicly ruled out single market re-entry during this parliament. That position will hold formally and shift operationally. Expect de facto alignment — on data, chemicals, product standards, and critical infrastructure procurement — to advance materially within 18 months, regardless of de jure status. By October 2027, the operational distinction between UK and EU regulatory regimes for sovereign technology procurement will be substantially smaller than it is today.

The bottom line

The board paper being written this week, in the commuter-belt office of a senior risk director somewhere between Manchester and London, has a choice. It can describe the events of the last fortnight as an oil shock plus three unrelated technology stories. Or it can describe them as a single inversion — the moment the post-war bargain stopped underwriting prosperity and began pricing it. The first description is comfortable and wrong. The second is uncomfortable and accurate.

The Control Layer exists to translate a Strait of Hormuz headline into a page-three line item in your Audit and Risk Committee paper, in a voice a board will quote — because nobody else on Substack is doing that translation, and the quarter that has begun is the one in which boards either commission that translation or discover, in the next quarter’s loss provision, that they should have.

Subscribe to receive Part 2 — The metal floor: aluminium, helium, and the hidden inputs to the sovereign technology stack — and the rest of the Four Chokepoints series.

Subscribe now

References

[1]: Reuters. *“Gulf aluminium output crippled as Iranian strikes hit EGA and Alba facilities.”* 11 April 2026.

[2]: U.S. Geological Survey. *“Mineral Commodity Summaries 2026 — Helium.”* January 2026.

Linde plc. *“Annual Report 2025 — Helium Supply Chain Disclosures.”*

[3]: U.S. House of Representatives. *Multilateral Alignment of Technology Controls on Hardware Act*, H.R. 8170, 119th Congress (2026).

[4]: HM Government and Government of the French Republic, joint communiqué. *“Franco-British Defensive Maritime Mission to the Persian Gulf.”* 10 April 2026.

[5]: ASML Holding N.V. *“Annual Report 2025 — Geographic Revenue Disclosure.”*

[6]: U.S. Bureau of Industry and Security. *“Export Licensing Performance Report Q1 2026.”*

[7]: Government of Spain, Office of the Prime Minister. *“State Visit to the People’s Republic of China, 13–15 April 2026.”*

[8]: Eurostat. *“EU trade in goods with China — 2025 annual data.”* Released 10 April 2026.

[9]: Helen Thompson. *Disorder: Hard Times in the 21st Century*. Oxford University Press, 2022.

[10]: Adam Tooze. *Chartbook* (Substack). See in particular Chartbook #321 on the structural polycrisis framework.

[11]: Drewry Shipping Consultants. *“Container Shipping Outlook April 2026.”* Cited in *Lloyd’s List*, 11 April 2026.

[12]: Noah Smith. *Noahpinion* (Substack). *“Derisking, not decoupling: the West’s actual China strategy.”*

Author

Amer Altaf is Founder and CEO of Arkava, a UK and European sovereign AI agentic automation business, and Managing Editor of The Control Layer, the publication where he tracks the convergence of cyber security, technology sovereignty, and geopolitics. A techUK member, he contributes to industry engagement on UK technology sovereignty policy. His January 2026 analysis of Airbus’s migration off Google and Microsoft argued European cloud sovereignty had moved from policy aspiration to active corporate posture — six months before the Hormuz blockade made the argument unavoidable. He is currently writing on cloud security for Oxford University Press’s Expert Essentials series.

In just weeks of testing, Mythos Preview has autonomously identified thousands of zero-day vulnerabilities in every major operating system and every major web browser — including a 27-year-old flaw in OpenBSD, a 16-year-old bug in FFmpeg that automated testing missed five million times, and a chained Linux kernel exploit that escalates to full machine control.

The 12 launch partners — AWS, Apple, Microsoft, Google, Cisco, CrowdStrike, Palo Alto Networks, Broadcom, NVIDIA, JPMorganChase, the Linux Foundation, and Anthropic — will use the model exclusively for defensive security work. Anthropic is committing $100 million in usage credits and $4 million to open-source security organisations.

In this solo episode, I break down what Mythos Preview can actually do, why the defensive case is strong, and why the dual-use problem — the same model that finds vulnerabilities can exploit them — cannot be engineered away.

Then I ask the question almost no one else covering this story is asking: why are all 12 launch partners US-headquartered? What does it mean when the most powerful defensive cybersecurity tool ever created is exclusively in the hands of American companies, subject to US government engagement, with no mention of the UK’s NCSC, the EU’s ENISA, or any non-US government body?

What I cover:

• Claude Mythos Preview’s capabilities — and why this is a step change, not an incremental improvement

• The defensive case: $100M in credits, open-source funding, and a coalition that touches most of the world’s software infrastructure

• The dual-use tension: Mythos develops working exploits autonomously, without human steering

• The sovereignty question: all 12 partners are US-headquartered, and the implications for UK and European defenders are significant

• Five things to watch over the coming months — from the 90-day report to the UK’s Cyber Security and Resilience Bill

This episode is for:

• CISOs and security leaders assessing what AI-augmented threats mean for their organisations

• CTOs and engineers building on infrastructure maintained by Glasswing partners

• Policymakers writing cybersecurity legislation in a world that just changed

• Anyone who believes the geography of AI capability is a strategic question, not a technical footnote

Read the full analysis:

Listen and subscribe:

• YouTube:

• Apple Podcasts:

• Spotify:

• Substack:

The Control Layer is hosted by Amer Altaf, founder and CEO of Arkava, and publishes weekly.

Sponsored by Arkava — Trusted Intelligence, Tangible Impact.

Yesterday, Anthropic did something no major AI lab has done before: it announced a frontier model — Claude Mythos Preview — and simultaneously declared it would not be making it publicly available. The reason? The model is so proficient at finding and exploiting software vulnerabilities that releasing it could, in the company’s own assessment, pose a serious threat to global cybersecurity.

Instead, Anthropic launched Project Glasswing — a coalition of 12 launch partners, plus roughly 40 additional organisations, who will use Mythos Preview exclusively for defensive security work. The stated goal is to find and fix vulnerabilities in the world’s most critical software before attackers can exploit them.

Join The Control Layer for weekly perspectives on AI, cybersecurity, and building technology that serves human purpose.

It is, by any measure, a significant moment. But it also raises questions that Anthropic’s announcement — carefully framed in the language of collective defence — leaves largely unanswered.

What Mythos Preview can actually do

The capabilities are striking. In just a few weeks of testing, Mythos Preview has autonomously identified thousands of zero-day vulnerabilities— flaws previously unknown to the software’s own developers — across every major operating system and every major web browser.

Three examples illustrate the scale of what we are talking about:

27-year-old vulnerability in OpenBSD, widely regarded as one of the most security-hardened operating systems in the world and used to run firewalls and critical infrastructure. The flaw allowed an attacker to remotely crash any machine running the OS simply by connecting to it. Twenty-seven years. Millions of lines of human review. Missed.

16-year-old vulnerability in FFmpeg, the video encoding library embedded in countless pieces of software globally. Automated testing tools had hit the relevant line of code five million times without catching the problem. Mythos Preview found it.

And perhaps most consequentially, the model autonomously discovered and chained together several vulnerabilities in the Linux kernel — the software running most of the world’s servers — to escalate from ordinary user access to complete machine control.

On the CyberGym benchmark for vulnerability reproduction, Mythos Preview scored 83.1%, compared to 66.6% for Anthropic’s previous best model, Claude Opus 4.6. On SWE-bench Verified, it reached 93.9%. These are not incremental improvements. They represent a step change in what AI systems can do to — and for — software security.

Anthropic’s own framing is unusually candid: frontier AI models have reached a level of coding capability where they can surpass all but the most skilled humans at finding and exploiting software vulnerabilities. And given the rate of AI progress, it will not be long before such capabilities proliferate.

The defensive case — and it is a strong one

The logic behind Project Glasswing is sound and, in many respects, overdue.

The global cost of cybercrime is notoriously difficult to pin down, but credible estimates place it between $500 billion and $1.5 trillion annually, depending on methodology. The attack surface is vast and growing: banking systems, healthcare records, logistics networks, energy grids, and the open-source software that underpins virtually all of them.

Until now, finding serious vulnerabilities has required a level of expertise held by a tiny number of people. AI models like Mythos Preview could fundamentally shift the economics of defence. If you can scan a codebase of millions of lines and surface flaws that survived decades of human review, you can fix them before they are exploited. That is genuinely transformative.

The open-source dimension is particularly important. The Linux Foundation’s Jim Zemlin put it well: open-source maintainers — whose software underpins much of the world’s critical infrastructure — have historically been left to figure out security on their own. Anthropic’s commitment of $2.5 million to Alpha-Omega and OpenSSF through the Linux Foundation, and $1.5 million to the Apache Software Foundation, is a concrete step toward changing that. If maintainers of critical open-source projects can access Mythos-class scanning, the downstream security benefits could be enormous.

The broader coalition makes strategic sense, too. AWS, Microsoft, Google, Apple, Cisco, CrowdStrike, Palo Alto Networks, Broadcom, NVIDIA, JPMorganChase — these organisations collectively touch a vast proportion of the world’s software and infrastructure. If they can systematically scan and patch their foundational systems, the aggregate reduction in attack surface would be significant.

Anthropic has also committed $100 million in usage credits for Mythos Preview across the initiative. After the research preview, the model will be available to participants at $25/$125 per million input/output tokens via the Claude API, Amazon Bedrock, Google Cloud’s Vertex AI, and Microsoft Foundry. This is not a symbolic gesture. It is a serious resource commitment.

The dual-use problem no one can engineer away

Here is the tension at the heart of Project Glasswing, and it is one Anthropic is remarkably transparent about: the same model that finds vulnerabilities can exploit them.

Mythos Preview did not simply identify the OpenBSD flaw or the Linux kernel vulnerabilities. It developed working exploits — in many cases entirely autonomously, without human steering. It wrote a browser exploit chaining together four separate vulnerabilities. It found KASLR bypasses and race conditions in the Linux kernel and escalated privileges to root.

This is why Anthropic is not releasing the model publicly. It is also why the company has been in ongoing discussions with US government officials about Mythos Preview’s offensive and defensive cyber capabilities.

But acknowledging the dual-use problem is not the same as solving it. Anthropic’s stated plan is to develop safeguards that “detect and block the model’s most dangerous outputs” and launch them with an upcoming Claude Opus model. Security professionals affected by those safeguards will be able to apply for an upcoming Cyber Verification Programme.

The question is whether safeguards can ever be robust enough. The history of AI safety measures suggests a pattern: capabilities advance faster than controls. And the specific challenge with cybersecurity capabilities is that the line between legitimate defensive research and offensive exploitation is often a matter of intent, not technique.

CrowdStrike’s CTO Elia Zaitsev captured the urgency well: “The window between a vulnerability being discovered and being exploited by an adversary has collapsed — what once took months now happens in minutes with AI.” That compression does not just apply to known vulnerabilities. It applies to the proliferation of AI models capable of finding new ones.

DARPA recognised this trajectory a decade ago with the original Cyber Grand Challenge in 2016. The AI Cyber Challenge in 2025 saw competing systems identify 86% of synthetic vulnerabilities, up from 37% at semifinals. Mythos Preview represents another leap beyond even that. The direction is clear; the speed is what has changed.

Twelve partners. All American. That is not a detail — it is the story.

Now consider the composition of Project Glasswing.

The 12 launch partners are: Amazon Web Services, Anthropic, Apple, Broadcom, Cisco, CrowdStrike, Google, JPMorganChase, the Linux Foundation, Microsoft, NVIDIA, and Palo Alto Networks.

Every single one is US-headquartered.

The 40+ additional organisations given access are described as those that “build or maintain critical software infrastructure.” The announcement does not name them. Anthropic frames the initiative in the language of global defence — “securing the world’s most critical software” — and references “democratic states” and “the United States and its allies.”

But the practical reality is that the most powerful defensive cybersecurity tool ever created is, at launch, exclusively in the hands of American companies and subject to US government engagement. Anthropic has been in “ongoing discussions with US government officials” about Mythos Preview. There is no mention of equivalent engagement with the UK’s National Cyber Security Centre, the EU Agency for Cybersecurity (ENISA), or any non-US government body.

This matters for several reasons.

First, critical national infrastructure does not respect corporate headquarters. The UK’s National Health Service, its energy grid, its financial services sector, and its defence supply chain all run on software maintained by the organisations in Project Glasswing. When Mythos Preview finds a vulnerability in the Linux kernel or in a Cisco product, that vulnerability affects British systems as much as American ones. The question is whether UK defenders will have the same access to fix it, at the same speed.

Second, Anthropic’s pricing model after the research preview — $25/$125 per million tokens — means that even when access is theoretically available, cost and availability will be determined by a US company operating under US jurisdiction. For organisations in regulated UK sectors — defence, financial services, healthcare — the question of where their vulnerability data is processed, and under whose legal jurisdiction, is not academic.

Third, and most fundamentally, this is a pattern. The most consequential AI capabilities — whether in language, reasoning, or now cybersecurity — are being developed by US companies, deployed first to US partners, and shaped by US government engagement. The UK’s position as a close ally does not automatically translate into equivalent access, timing, or influence over how these tools are governed.

None of this is a criticism of Anthropic’s decision to restrict Mythos Preview. On the contrary — the decision not to release it publicly is responsible, perhaps even courageous, given the commercial pressure to ship. But responsible AI development and equitable access to AI-powered defence are two different things. Project Glasswing addresses the first. It does not yet address the second.

What to watch

Project Glasswing is the beginning of something, not the end. Several things are worth tracking over the coming months:

The 90-day public report. Anthropic has committed to reporting publicly within 90 days on what the initiative has learned, including vulnerabilities fixed and improvements made. The quality and transparency of that report will tell us a great deal about whether this is a genuine public good or a carefully managed preview programme.

Non-US participation. Whether and how Anthropic extends Glasswing access to non-US governments, CERTs, and critical infrastructure operators. The announcement mentions an aspiration for “an independent, third-party body” to host continued work. Who sits on that body, and where it is domiciled, will matter.

The safeguards question. Anthropic’s plan to launch new safeguards with an upcoming Claude Opus model is the critical dependency for any broader availability of Mythos-class capabilities. If those safeguards prove robust, it opens the door to wider defensive use. If they do not, the argument for permanent restriction strengthens — and so does the asymmetry between those who have access and those who do not.

Open-source uptake. The $4 million in direct donations and the Claude for Open Source programme are the most democratising elements of the announcement. Whether maintainers of critical open-source projects actually adopt Mythos-class scanning — and whether the tools are accessible enough for under-resourced teams — will determine whether this narrows or widens the security gap.

Government response. The UK government’s forthcoming Cyber Security and Resilience Bill, and the EU’s implementation of the NIS2 Directive, will both need to reckon with a world in which AI systems can find and exploit vulnerabilities faster than human defenders can patch them. Whether policymakers are prepared for that shift is an open question.

The bottom line

Project Glasswing is the most significant cybersecurity announcement of 2026 so far, and arguably the clearest demonstration yet that frontier AI has crossed a threshold in offensive and defensive cyber capability.

Anthropic deserves credit for the transparency of its approach and the scale of its commitment. The decision to restrict Mythos Preview, to build a coalition rather than race to market, and to fund open-source security work — these are all the right instincts.

But the announcement also crystallises a deeper tension in the AI era: the tools we most urgently need for collective defence are being built, controlled, and governed by a remarkably small number of actors in a single jurisdiction. For the UK, for Europe, and for every other democracy that depends on the same software infrastructure, the question is not whether Project Glasswing is a good idea. It is. The question is whether we are content to be downstream beneficiaries of American AI capability — or whether we intend to be active participants in shaping how that capability is deployed.

That is not a question Anthropic can answer. It is one for us.

Amer Altaf is Founder & CEO of Arkava, a sovereign AI agentic automation business in the United Kingdom and Europe, and Managing Editor of The Control Layer.

Sources and further reading:

- [Project Glasswing — Anthropic]

- [Claude Mythos Preview — Anthropic Frontier Red Team]

- Anthropic limits Mythos AI rollout over fears hackers could use model for cyberattacks — CNBC

- [Anthropic is giving some firms early access to Claude Mythos to bolster cybersecurity defenses — Fortune]

- [Anthropic says its most powerful AI cyber model is too dangerous to release publicly — VentureBeat]

- [Tech giants launch AI-powered Project Glasswing — CyberScoop]

- [CrowdStrike founding member of Anthropic Mythos initiative]

- [Project Glasswing: Tech giants unite to fix AI-found software risks — Interesting Engineering]

- [DARPA AI Cyber Challenge results — DARPA]

- [Estimating global yearly cybercrime damage costs — Governance.ai]

Why it matters: Three quarters of enterprise AI investments fail to produce measurable value. New cross-sector evidence from Broadridge, Gartner, and verified case studies across energy, logistics, and financial services reveals a consistent cause. The problem is not the AI. It is the absence of an orchestration layer connecting discrete technologies into coordinated, accountable workflows. Organisations that close this gap are reporting 50% cycle-time reductions, millions in annual savings, and production gains of 10–20%.

Join The Control Layer for weekly perspectives on AI, cybersecurity, and building technology that serves human purpose.

Why is the gap between AI spending and AI value widening?

Something peculiar is happening in enterprise technology. Spending on AI has never been higher. Adoption has never been faster. And yet, the majority of AI investments still fail to return measurable business value.

The Broadridge 2026 Digital Transformation Study, surveying more than 900 financial services technology and operations leaders globally, documents a striking acceleration: 80% of firms now report active AI use, up from 31% in 2025 [1]. Some 26% are already deploying agentic AI, with more than half of those deployments moving beyond pilot programmes into operational use [1]. One US-based asset manager reported saving approximately 350 hours per month in its KYC function alone through agentic AI [1].

Yet look beneath the headline adoption figures and a more uncomfortable picture emerges. Forty-three percent of firms believe they will need to rebuild their entire technology stack to succeed in the age of AI [1]. Sixty-five percent have no formal mandate or incentives to use AI. Thirty-seven percent cite talent shortages as a barrier to scaling agentic systems [1].

The pattern is consistent across sectors. Investment is accelerating. Value realisation is not keeping pace. According to McKinsey, while the economic potential of AI is enormous, most organisations struggle to move from pilot to production at scale. The question is: why?

What is orchestration and why does it matter?

The answer, increasingly supported by cross-sector evidence, is orchestration. Not the AI models themselves. Not the data. Not the algorithms. The layer that sits between all of them, coordinating decisions, enforcing governance, and connecting discrete technologies into workflows that actually produce outcomes.

Service Orchestration and Automation Platforms (SOAPs) have evolved from specialist IT scheduling tools into what Gartner now recognises as a distinct, critical market category [2]. Redwood Software, named a Leader in Gartner’s Magic Quadrant for SOAPs for two consecutive years, frames the challenge plainly: the most competitive enterprises are no longer differentiating on the smartness of their algorithms, but on how effectively their orchestration platforms connect insight to decision to execution [2].

Think of it this way. A hospital might have an excellent diagnostic AI, a robust electronic patient record, and an efficient scheduling system. Without orchestration, each operates in isolation. The diagnostic insight does not automatically trigger the right appointment. The appointment does not automatically update the patient record. The record does not automatically inform the discharge process. Each handoff introduces delay, error, and cost.

Orchestration eliminates those handoffs. It turns a collection of capable tools into a coordinated system. The UK’s National Cyber Security Centre has long advocated integrated, defence-in-depth approaches to technology architecture, and orchestration applies the same principle to operational AI.

The enterprises realising extraordinary AI returns are not the ones with the smartest algorithms. They are the ones with orchestration platforms that connect discrete technologies into coordinated workflows.

What does orchestration look like in practice?

Financial services: from pilots to production

The Broadridge data confirms financial services is leading the orchestration shift. Among firms deploying agentic AI, adoption is most advanced at large institutions. Nearly one third of firms managing more than $250 billion in assets report active agentic AI deployments [1]. These are production systems handling underwriting, compliance interpretation, and customer onboarding.

The critical enabler is orchestration. Autonomous agents handling KYC checks or loan decisioning only work at scale when an orchestration layer governs their behaviour, ensures compliance with frameworks such as the EU AI Act and UK GDPR, maintains audit trails, and integrates with legacy ERP systems without disruption.

Energy: process mining before automation

OMV Petrom, Europe’s largest oil and gas producer, provides a compelling case study in orchestration-first transformation. The company’s S4Strive programme consolidated 170 legal entities under a single SAP S/4HANA platform, integrating business process management frameworks, process mining analytics, and robotic process automation for financial workflows [3].

The results were substantial: a 50% reduction in invoice-to-payment cycle times, more than 225,000 labour hours saved annually, and cost reductions exceeding €8 million [3]. IBM, the implementation partner, confirmed the consolidation of 170 company codes into a single instance [4].

The critical lesson here is sequencing. OMV Petrom did not deploy RPA first and hope for the best. The organisation standardised processes through BPM frameworks, then used process mining to identify bottlenecks and deviations, and only then applied automation to the right tasks in the right order. Orchestration was the design principle, not an afterthought. This mirrors the approach recommended by the NCSC’s Cyber Assessment Framework: understand your current state before implementing controls.

Manufacturing: scheduling as the orchestration centre

BDO’s 2026 Manufacturing Industry Predictions highlight the growing role of predictive modelling and analytics tools for identifying supply chain risks, projecting stock-outs, and adjusting production schedules dynamically [5]. Early adopters report 10–20% production gains and 7–20% productivity improvements from operational AI tied to sensors, schedules, and routings [6].

Scheduling systems are emerging as orchestration centres that translate supply chain anomalies, maintenance signals, and quality alerts into coordinated production decisions. The Deloitte 2025 Smart Manufacturing Survey found that 29% of manufacturers deploy traditional AI/ML for operations, compared to 24% for GenAI, with operational AI delivering significantly higher ROI [6].

Logistics: RPA at scale requires orchestration discipline

Raben Group, a Dutch-based logistics provider managing 1.6 million square metres of warehouse space and nearly 12,000 employees, demonstrates what happens when RPA is deployed with orchestration rigour. The company has deployed more than 200 RPA automations, saving 78,815 employee workdays and over €6 million annually [7].

One telling detail: the average time for manual creation of spot customer offers was 15 minutes. After automation, it dropped to an average of 21 seconds, with over 99% sent to customers within two minutes [8]. Raben’s success correlates directly with its orchestration approach: a proprietary process management platform, a dedicated Centre of Excellence, and systematic identification of high-volume processes before deploying automation [8].

How does the UK public sector approach orchestration?

The UK Government Digital Service Roadmap 2026–2030 signals that orchestration principles are entering public sector strategy. GOV.UK One Login has been used by over 13 million people to access more than 120 government services [9]. The roadmap mandates API-first architecture across public sector organisations, with a target of 1 in 10 civil servants in digital and technology roles by 2030 [9].

For UK suppliers and technology consultancies, this represents both opportunity and obligation. The Digital Commercial Centre of Excellence projects £1.2 billion in annual savings through standardised procurement [9]. Vendors unable to meet government data residency, Cyber Essentials Plus, or ISO 27001 compliance requirements face exclusion from this growing market.

Analysis: three lessons for UK organisations

First, orchestration is not optional infrastructure. Point solutions without orchestration create silos that degrade AI value realisation and increase total cost of ownership. An organisation deploying RPA here, a predictive model there, and an AI chatbot somewhere else without connecting them is building complexity rather than capability.

Second, process standardisation must precede automation. OMV Petrom’s success was built on process mining and BPM standardisation before any RPA was deployed. Raben Group’s results correlate with systematic process identification and governance. The lesson is clear: automating a broken process produces a faster broken process.

Third, governance and orchestration are converging. As agentic AI moves into production, orchestration platforms are becoming the enforcement layer for AI governance under frameworks like the EU AI Act and the UK’s Data Use and Access Bill. They define boundaries for autonomous action, maintain audit trails, and ensure compliance. With 37% of firms citing talent gaps [1], governance cannot rely on human oversight alone. It must be embedded in the orchestration layer.

How Arkava helps enterprises close the orchestration gap

Arkava exists precisely because mid-market UK organisations face the orchestration gap most acutely. Enterprise platforms like Palantir and Workato serve organisations with six-figure technology budgets and dedicated integration teams. The mid-market, typically £50M–£500M revenue, sits in a gap: too complex for off-the-shelf tools, too resource-constrained for enterprise platforms.

Arkava bridges this gap through three interconnected services. Arkava Advisory provides strategic consulting that begins with process mapping and existing technology assessment, identifying where disconnected AI investments are leaking value before recommending solutions. This mirrors the OMV Petrom sequencing principle: understand the landscape, standardise the processes, and only then deploy automation.

Arkava Ignite delivers AI transformation assessments that evaluate an organisation’s readiness for orchestrated automation, covering data quality, process maturity, governance frameworks, and integration architecture. The assessment produces a prioritised roadmap calibrated to the organisation’s actual capabilities rather than aspirational targets.

Arkava Spark, the subscription automation platform, provides the orchestration layer itself, built on Industry-leading workflow orchestration technology and deployed within UK sovereign infrastructure. This means data remains under UK jurisdiction, meeting UK GDPR, DPA 2018, and defence-sector compliance requirements without exposing organisations to CLOUD Act risks associated with US-headquartered cloud providers.

The outcome-based pricing model aligns Arkava’s incentives with client results. Rather than billing for implementation hours regardless of outcome, Arkava ties success to measurable value realisation. This is the difference between buying technology and buying outcomes.

For organisations recognising the orchestration gap in their own operations, a practical first step is the Arkava Ignite Readiness Assessment: a structured evaluation of current AI investments, process maturity, and integration architecture that identifies the highest-value orchestration opportunities. Contact Arkava to discuss your organisation’s specific situation.

Risks and constraints

Integration complexity is real. Consolidating legacy systems under an orchestration layer is a multi-year undertaking, as OMV Petrom’s experience demonstrates. For mid-market organisations with smaller IT teams, external expertise is often essential.

Vendor lock-in risk. SOAP platforms are becoming critical infrastructure. Organisations should evaluate data portability, API openness, and exit costs before committing. UK sovereignty considerations add another dimension: where is your orchestration data stored, and under whose jurisdiction?

The talent gap applies here too. Orchestration requires people who understand both technology and business processes. The UK Tech Workforce shortage means this hybrid skillset remains scarce.

Measurement remains imperfect. Attribution is complex when isolating orchestration value from underlying technology value. Organisations should expect ROI measurement to be iterative.

What should you do next?

For boards and executives: Commission an orchestration capability assessment. The question is not “do we have enough AI?” It is “are our AI investments connected to each other and to business outcomes?”

For technical leaders: Evaluate readiness for orchestration platforms. Prioritise process mining and standardisation before deploying additional automation. Establish governance frameworks for agentic AI embedded in the orchestration layer, aligned with ICO guidance and sector-specific regulations.

For mid-market organisations: Start with your highest-value, highest-volume processes. Apply the OMV Petrom sequence: standardise first, mine for bottlenecks second, automate third. Even connecting two previously siloed tools into a coordinated workflow can deliver measurable returns within months.

Disclaimer: This article represents analysis based on publicly available information as of March 2026. It does not constitute legal, financial, or professional advice.

*If your organisation needs support connecting AI investments to measurable outcomes, Arkava helps mid-market enterprises turn technology spending into business results through sovereign orchestration and governance frameworks.*

References

[1] Broadridge Financial Solutions. “2026 Digital Transformation & Next-Gen Technology Study.” *Broadridge*, 25 February 2026.

[2] Redwood Software. “AI and Automation Trends 2026: From Efficiency to Enterprise Resilience.” *Redwood*, 28 January 2026.

[3] Ionescu, V. & Popescu, A. “Digital integration for sustainable competitiveness: the role of BPM, process mining and ERP in accounting automation.” *Applied Studies in Agribusiness and Commerce*, 18(1), 31 December 2025.

[4] IBM. “OMV prepares for sustainable growth.” *IBM Case Studies*.

[5] BDO USA. “2026 Manufacturing Industry Predictions.” *BDO Insights*, 16 March 2026.

[6] Phantasma Global. “AI and Automation in Manufacturing: Trends, Challenges & High ROI Use Cases in 2026.” *Phantasma Global*.

[7] Supply Chain Brain. “How Robotic Process Automation Helps Reduce Business Costs.” *Supply Chain Brain*, 10 September 2024.

[8] UiPath. “Automation saves Raben Group €6 million a year.” *UiPath Case Studies*.

[9] Government Digital Service. “Our roadmap for modern digital government.” *GDS Blog*, 20 January 2026.

Last week? Yesterday? This morning?

Now imagine you didn’t do any of it. Your AI assistant found the product, compared the prices, checked the reviews, and completed the purchase. All while you were drinking your morning coffee.

That world isn’t coming. It’s already here. And it raises a question that very few people are asking, but everyone should be: who controls the agent?

When an AI is shopping on your behalf — spending your money, handing over your data — who’s making sure it’s safe? Who’s making sure it’s even really you?

That’s the question I put to Andrew Dunbar in the very first episode of The Control Layer.

Andrew is the Chief Information Security Officer at Shopify. He joined as their first security hire in 2012 and has spent 14 years building the trust infrastructure behind a platform that now powers commerce for over 875 million buyers a year — roughly one in six people on the internet.

What We Cover

The new perimeter. The browser used to be the security boundary. The padlock icon, the HTTPS warning, the cookie consent pop-up — all of it was designed to keep humans in the loop. In the agentic world, the browser is gone. So what replaces it?

The Universal Commerce Protocol. In January 2026, Shopify and Google co-launched UCP — an open protocol that defines how AI agents, merchants, credential providers, and payment processors should interact. Andrew walks through the four roles, why it’s open rather than proprietary, and how cryptographic signing means a compromised agent can’t complete a fraudulent transaction even if the platform or the merchant has been breached.

The mobile OS analogy. Andrew makes a point that stuck with me: mobile operating systems got to learn from every mistake desktop ever made. Containerisation, permissioning, isolation — all built in from day one. Agentic AI is that same generational leap. We have a chance to build trust in from the start rather than bolting it on later.

The 15x signal. Shopify has seen a 15x year-over-year increase in agentic shopping leading to actual purchases. This isn’t theoretical. The shift is happening now, and the security architecture needs to move just as fast.

What every CISO should be doing today. Andrew’s advice is direct: deploy phishing-resistant MFA everywhere, audit how your vendors are using AI, and start instrumenting decisions — not just infrastructure. Chain-of-thought logging gives you something you’ve never had before: the ability to ask an agent what were you thinking when you did that?

This was a conversation I’ve wanted to have for months, and Andrew delivered. Whether you’re a founder, a security professional, a product builder, or someone who’s just started letting AI do your shopping — this episode is worth your time.

Listen Elsewhere

🎧 Apple Podcasts —

🎧 Spotify —

📺 YouTube —

If this is the kind of conversation you want more of — AI, cybersecurity, trust, and the people building the systems that sit between you and the machine — subscribe to The Control Layer. New episodes dropping regularly.

— Amer Altaf, Founder & CEO, Arkava

Why it matters: Frontier AI just proved it can find critical software vulnerabilities faster than human security researchers. At the same time, criminal groups are weaponising the same technology for autonomous attacks at machine speed. The organisations that harness AI for defence first will define the next era of cybersecurity. Those that wait will be defending yesterday’s perimeter against tomorrow’s threats.

Join The Control Layer for weekly perspectives on AI, cybersecurity, and building technology that serves human purpose.

The Discovery That Changed the Conversation

On 6 March 2026, Anthropic announced that Claude Opus 4.6 had identified 22 critical vulnerabilities in Firefox 148, including 14 classified as high severity [3]. This was not a theoretical exercise or a controlled benchmark. It was a structured security research engagement that validated Mozilla’s patch cycle and demonstrated something the cybersecurity community has debated for years: frontier AI models can perform productive vulnerability research at a speed and scale that human teams simply cannot match.

The significance extends beyond the headline number. Traditional vulnerability research relies on skilled humans spending weeks or months fuzzing code, reviewing logic flows, and testing edge cases. Claude completed its analysis using extended context capabilities processing up to one million tokens, effectively reading and reasoning across vast codebases in a single pass [3]. The model’s adaptive thinking and reduced hallucination rates (80% reduction compared to extended thinking baselines) meant its findings were not just numerous but reliable enough to action [6].

For UK mid-market organisations that cannot afford dedicated red teams or bug bounty programmes costing six figures annually, this development is not academic. It signals a future where AI-augmented security tooling becomes accessible, practical, and potentially transformative.

The Other Side of the Coin

Here is where the story gets uncomfortable. While Anthropic was demonstrating AI’s defensive potential, criminal ecosystems were scaling offensive AI capabilities at an alarming rate.

Flashpoint’s 2026 Global Threat Intelligence Report documents what it calls an era of “total convergence” in cybercrime [4]. The numbers demand attention: illicit discussions about AI and machine learning on underground forums surged 1,500% between November and December 2025, jumping from approximately 360,000 to 6 million messages [4]. This is not idle chatter. These communities are building, sharing, and refining autonomous attack frameworks that execute complete attack sequences without human intervention. Reconnaissance, phishing, credential testing, infrastructure rotation — all orchestrated by AI agents operating at machine speed.

The infostealer epidemic provides the fuel. In 2025 alone, 11.1 million devices were infected with credential-harvesting malware, yielding 3.3 billion stolen credentials and cloud tokens [4]. Ransomware incidents climbed 53% year-over-year, with 87% attributed to Ransomware-as-a-Service groups [4][15]. The tactical evolution is stark: attackers now prefer logging in with stolen credentials over breaking in through technical exploits. When your adversary already has the keys, your firewall is decorative.

Perhaps most concerning is the compression of exploitation timelines. The window between vulnerability disclosure and active exploitation has collapsed from weeks to hours, with high-impact flaws now weaponised within 4 to 8 hours of CVE publication [4]. This means traditional patch management cycles — where organisations review, test, and deploy fixes over days or weeks — are increasingly inadequate against adversaries operating at AI-accelerated tempo.

The Arms Race Nobody Can Opt Out Of

The cybersecurity landscape has split into two speeds: AI-accelerated offence operating in hours, and human-paced defence operating in weeks. That gap is where breaches live.

Think of it like this. Imagine a chess match where one player has a grandmaster AI whispering moves in real time, and the other is consulting a strategy book from 2019. That is roughly the asymmetry facing organisations that have not yet integrated AI into their defensive operations.

IBM X-Force researchers discovered over 300,000 ChatGPT credentials listed for sale on dark web marketplaces in 2025 [17]. Open-source AI agent platforms, designed for legitimate automation, are being repurposed as credential harvesting infrastructure. When compromised, these agent systems become goldmines of authentication tokens, API keys, and session data [17].

The World Economic Forum’s February 2026 threat assessment reinforces the picture: 87% of security leaders now identify AI-related vulnerabilities as the fastest-growing risk category, yet only 64% have adopted post-deployment AI security assessments [18]. That 23-point gap between awareness and action is precisely where organisational risk accumulates.

This is not a future problem. It is a current one accelerating weekly.

What Defenders Can Actually Do

The Firefox discovery offers a blueprint, not just a headline. It demonstrates three principles that UK organisations of any size can begin applying.

Principle 1: AI as Force Multiplier, Not Replacement. Claude did not replace Mozilla’s security team. It augmented their capability, identifying flaws that could then be verified, prioritised, and patched by human engineers. The model brought scale and speed. Humans brought context, judgement, and accountability. This partnership model is how AI-augmented security will work in practice — not autonomous systems making unilateral decisions about your infrastructure, but intelligent tools that surface risks for human decision-makers to act upon.

Principle 2: Defensive AI Must Match Offensive Tempo. If attackers exploit vulnerabilities within hours of disclosure, defenders need detection and response capabilities operating on the same timescale. AI-powered vulnerability scanning, anomaly detection, and automated triage can compress the defender’s loop from days to minutes. The technology exists. The implementation gap is organisational, not technical.

Principle 3: Identity Is the New Perimeter. With credential theft dominating the attack landscape and ransomware groups preferring to log in rather than break in [4][15], traditional network security is necessary but insufficient. Multi-factor authentication, privileged access management, single sign-on hardening, and credential monitoring are no longer optional security enhancements. They are baseline survival requirements. AI-powered behavioural analytics can detect anomalous login patterns — a London-based finance director’s credentials used from an unfamiliar jurisdiction at 3am — in ways that static rules cannot.

Analysis

In my view, the Firefox vulnerability discovery represents an inflection point, though perhaps not the one most commentary has focused on.

The real story is not that AI found bugs. Security researchers have been using automated tools for decades. The story is the reliability and depth of what a general-purpose frontier model achieved without being specifically trained for security research. Claude Opus 4.6 was not a purpose-built vulnerability scanner. It was a reasoning model applied to a security context [3]. This generality matters enormously because it means organisations do not need to wait for specialised security AI products to enter the market. They can begin integrating existing frontier models into their security workflows today.

The flip side is sobering. If a commercially available model can find 22 critical flaws in a major browser, criminal groups with access to similar (or open-weight) models can do the same — and they are under no obligation to report their findings responsibly. DeepSeek V4, released as open-weight with one trillion parameters, provides capability that anyone can deploy without usage restrictions or ethical guardrails [1]. The democratisation of AI cuts both ways.

For UK mid-market organisations specifically, the strategic question is not whether to adopt AI-augmented security but how quickly they can do so without introducing new risks. Rushing to deploy AI tools without proper governance — clear access controls, credential isolation, activity logging, and incident escalation protocols — simply creates a different attack surface. The organisations that will navigate this transition successfully are those that treat AI security adoption as a governance challenge, not merely a technology procurement exercise.

Risks and Constraints

Hallucination risk remains real. While Claude Opus 4.6 demonstrated an 80% reduction in hallucination rates, that still means false positives will occur [6]. Security teams acting on AI-generated vulnerability reports without human verification risk wasting resources on phantom threats or, worse, introducing instability through unnecessary patches.

Asymmetric access to capability. Large enterprises can integrate frontier AI models into sophisticated security operations centres. Mid-market organisations with smaller teams and tighter budgets may struggle to operationalise these tools without external support or simplified platforms.

Regulatory ambiguity around AI in security contexts. The UK’s Data Use and Access Act 2025 permits automated decision-making with safeguards, but the boundaries of acceptable AI-driven security responses (automated blocking, credential revocation, incident escalation) remain unclear in practice [5]. Organisations deploying AI defensively should document their governance frameworks now, before enforcement precedents crystallise.

Open-weight model proliferation. DeepSeek V4’s open-weight release provides adversaries with frontier-class capability without commercial oversight [1]. Defensive strategies must assume adversaries have access to equivalent or superior AI tools.

What to Do Next

For boards and executives: Commission an AI security readiness assessment before Q3 2026. Understand where your organisation’s defensive tempo falls relative to the current threat landscape. Ask your CISO one question: if a critical vulnerability is published tomorrow morning, how many hours until we are patched? If the answer exceeds 24 hours, your patch management process needs AI augmentation.

For technical leaders: Begin piloting frontier AI models for internal vulnerability scanning and code review. Start with non-critical systems to validate reliability and build confidence. Simultaneously, audit your identity security stack — MFA coverage, PAM deployment, SSO configuration, credential monitoring. The shift from perimeter to identity-centric security is not optional.

For mid-market organisations: You do not need a dedicated AI security team to benefit from these developments. Evaluate managed security service providers that are integrating AI-powered threat detection. Prioritise identity hardening over network expansion. And establish clear governance for any AI tools touching your security infrastructure — access controls, logging, human oversight requirements, and incident response protocols.

Disclaimer: This article represents analysis based on publicly available information as of March 2026. It does not constitute legal, financial, or professional advice.

If your organisation needs support implementing AI-augmented security governance, Arkava helps mid-market enterprises turn AI investment into measurable security outcomes.

References

[1] Reuters. “DeepSeek to Launch New AI Model.” Reuters, 9 January 2026. https://www.reuters.com/technology/deepseek-launch-new-ai-model-focused-coding-february-information-reports-2026-01-09/
[2] AI Governance Desk. “EU AI Act Enforcement 2026: A CCO’s Complete Compliance Roadmap.” 14 January 2026. https://aigovernancedesk.com/eu-ai-act-enforcement-2026-cco-roadmap/
[3] The Hacker News. “Anthropic Finds 22 Firefox Vulnerabilities Using Claude Opus 4.6 AI.” 6 March 2026. https://thehackernews.com/2026/03/anthropic-finds-22-firefox.html
[4] Flashpoint. 2026 Global Threat Intelligence Report: Era of Total Convergence in Cybercrime. March 2026. https://www.techradar.com/pro/security/in-2026-cybercrime-has-reached-a-point-of-total-convergence-new-research-claims-ai-attack
[5] Osborne Clarke. “Artificial Intelligence | UK Regulatory Outlook February 2026.” 8 February 2026. https://www.osborneclarke.com/insights/regulatory-outlook-february-2026-artificial-intelligence
[6] Blog.mean.ceo. “New AI Model Releases News | March 2026.” 28 February 2026. https://blog.mean.ceo/new-ai-model-releases-news-march-2026/
[15] Convergence Networks. “Top Cyber Threats for 2026.” 18 February 2026. https://convergencenetworks.com/blog/top-cyber-threats-for-2026/
[17] IBM Security. “Cybersecurity Trends 2026: X-Force Threat Analysis.” 10 March 2026. https://www.ibm.com/think/insights/more-2026-cyberthreat-trends
[18] World Economic Forum. “Cyber Threats to Watch in 2026.” 17 February 2026. https://www.weforum.org/stories/2026/02/2026-cyberthreats-to-watch-and-other-cybersecurity-news/

Why it matters: On 3 March 2026, the CEO of Anthropic sat down for the most consequential interview in AI history. Not a keynote. Not a blog post. A direct, unscripted conversation about why his company told the Pentagon no — and what happens next. Every sentence was chosen carefully. This is what he said, what he meant, and what it tells you about where AI governance actually stands.

This is a companion piece to Who Really Controls Your AI? Trump's Ban Threatens UK-EU Sovereignty, which covers the geopolitical and sovereignty implications of the ban.
Watch the full interview here:

Join The Control Layer for weekly perspectives on AI, cybersecurity, and building technology that serves human purpose.

The Setup: Establishing Credentials Before Drawing Lines

Amodei opens not with principle but with proof of loyalty. His first substantive answer is a roll call of pro-military credentials.

He establishes that Anthropic was the first company to deploy on the classified cloud, the first to build custom national security models, and that Claude operates across the intelligence community and military for cyber operations and combat support [1]. He frames this explicitly as patriotic duty, citing the need to defend against autocratic adversaries like China and Russia [1].

This is deliberate sequencing. Before he says the word “no,” he wants the audience to understand that this is not coming from a pacifist. It is not coming from someone who objects to military work on principle. It is coming from the company that has done more military AI work than anyone else in the industry.

The framing matters because it pre-empts the most obvious attack — the one President Trump made hours later when he accused Anthropic of putting American lives at risk. Amodei’s answer to that accusation was already embedded in his opening statement. We are not refusing to serve. We are refusing two specific things out of hundreds.

His estimate — 98% or 99% of use cases accepted— is almost certainly rounded for rhetorical effect. But the message is clear. This is a boundary dispute, not a philosophical objection to military AI.

Red Line One: Surveillance That Is Legal but Shouldn’t Be

Amodei’s surveillance argument is the more intellectually ambitious of the two, and he frames it with a specific mechanism rather than abstract principle.

He describes a pipeline: private companies collect citizen data through commercial activity. The government purchases that data legally. Before AI, this was functionally useless — no human team could analyse billions of data points across millions of people. AI changes the equation. Suddenly, legally purchased commercial data becomes a mass surveillance apparatus [1].

The critical phrase is “getting ahead of the law”. He is not arguing that mass surveillance violates current legislation. He is arguing the opposite — that it does not, and that this is the problem. The legal framework was designed for a world where this capability was technically impossible. Nobody wrote laws against it because nobody needed to.

This is a more sophisticated argument than it first appears. He is not asking the Pentagon to obey the law. He is asking the Pentagon to respect the intent behind the law — the democratic principle that citizens should not be subjected to mass analysis by their own government — even where the letter of the law has not caught up with the technology.

For anyone who has watched the debate around the UK’s Investigatory Powers Act or the EU’s position on bulk data collection, this argument has immediate resonance. The gap between what is technically possible, what is legally permitted, and what is democratically acceptable is widening in every jurisdiction. Amodei is pointing at that gap and saying: we will not help you exploit it.

Red Line Two: The Engineering Argument Against Lethal Autonomy

The weapons argument is structurally different. Where the surveillance objection is about values, the weapons objection is about engineering.

Amodei draws a careful line between partially autonomous weapons — the kind deployed in Ukraine and potentially relevant to Taiwan — and fully autonomous systems where weapons fire without any human involvement. He explicitly acknowledges that democratic nations may eventually need fully autonomous weapons to defend against adversaries who develop them first.

This concession is significant. He is not ruling out autonomous weapons forever. He is ruling them out now, for two stated reasons.

The first is blunt: the AI is not reliable enough. His exact framing — “anyone who has worked with AI models understands there is a basic unpredictability to them that, in a purely technical way, we have not solved”— is the most candid public admission of AI unreliability from any frontier lab CEO. He is not hedging with “sometimes” or “in certain edge cases.” He is stating a fundamental property of current AI systems.

The second is the oversight gap. If you have a large army of drones or robots operating without human oversight, without human soldiers making targeting decisions, the question of who is responsible for what they do becomes unanswerable. He argues that conversation has not happened yet, and deploying the technology before it does is reckless.

Read together, the two arguments create a logical sequence. The technology is not reliable enough. The oversight framework does not exist. Therefore deployment is premature. He is not saying never. He is saying not yet, and not without the conversation we have not had.

This matters for organisations well beyond defence. The same logic applies to any AI system making consequential autonomous decisions — in lending, hiring, medical triage, security access. If there is a basic unpredictability you have not solved, and the oversight framework does not exist, deployment is premature. The principle scales.

The Three-Day Ultimatum: Power, Not Negotiation

The middle section of the interview reveals the mechanics of how the standoff escalated — and it does not read like a negotiation.

Amodei describes a three-day window: agree to Pentagon terms or face designation as a supply chain risk under the Defense Production Act. During that window, the Pentagon sent language that appeared to accommodate Anthropic’s concerns. But the language was loaded with escape clauses — “if the Pentagon deems it appropriate” and “to do anything in line with laws”.

These phrases deserve attention. “If the Pentagon deems it appropriate” transfers all discretion to the party making the request. It is not a restriction; it is a permission slip written as one. “In line with laws” sounds reasonable until you remember that Amodei’s entire surveillance argument rests on the fact that mass surveillance is currently lawful. A commitment to “lawful use” explicitly permits the very thing Anthropic is objecting to.

The Pentagon’s public position, reiterated by spokesman Sean Parnell — “We only allow all lawful use” — confirms this reading. It is not a concession. It is a restatement of the status quo dressed as accommodation.

Amodei’s description of this exchange is measured but pointed. He does not accuse anyone of bad faith. He simply notes that the proposed terms “did not actually concede in any meaningful way”. The restraint is itself a rhetorical choice. He is letting the audience draw the conclusion.

“Retaliatory and Punitive”: Choosing Words Under Pressure

The most revealing moment comes when the interviewer presses Amodei on whether the Pentagon’s actions constitute an abuse of power.

He deflects the first time: “I would return to the idea that this is unprecedented”. When pressed — “But is it an abuse of power?” — he deflects again, noting that this designation has never been used against an American company [1]. He then adds that government statements made it “very clear” this was “retaliatory and punitive” [1].

Watch the language architecture here. He will not say “abuse of power.” That phrase has legal implications and would position Anthropic as making a constitutional claim. Instead, he uses “unprecedented,” “retaliatory,” and “punitive” — words that describe the government’s behaviour without invoking a specific legal framework.

This is a CEO who knows he is heading to court and is being careful not to lock himself into a legal theory during a television interview. The restraint is professional, but the message is unmistakable. They punished us for saying no.

He reinforces this when asked about formal notification. Anthropic has received nothing official — no designation letter, no formal action [1]. Everything has come through social media posts from the President and Secretary Hegseth. The implication: the most powerful government in the world is conducting industrial policy through tweets.

“We Will Be Fine”: Confidence or Performance?

The interview’s closing minutes are about survival. The interviewer asks directly: can Anthropic survive this?

Amodei’s answer is emphatic: “Not only survive it; we are going to be fine” [1]. He characterises the impact of the designation as “fairly small” and accuses the government of deliberately creating “fear, uncertainty, and doubt” [1].

This is corporate crisis communication at a high level. Whether the impact is actually small is debatable — losing all US government contracts and being blacklisted from the defence supply chain is not trivial. But Amodei’s job in this moment is not accuracy. It is confidence. Customers, employees, and investors are watching. Any hint of vulnerability accelerates the exodus.

His repeated emphasis on continuity — offering to maintain services during a transition, supporting warfighters, helping off-board to a competitor — serves a dual purpose. It positions Anthropic as the responsible party (we are trying to help; they are being difficult). And it creates a record that may prove useful in court: we did everything reasonable to mitigate the disruption they caused.

What the Interview Does Not Say

What is absent from Amodei’s responses tells you as much as what is present.

He never mentions OpenAI or Grok by name. He never discusses the competitive dynamics of rivals stepping into the gap Anthropic left. This is almost certainly deliberate. Naming competitors draws comparison and concedes that alternatives exist.

He never discusses Anthropic’s commercial customers. No mention of enterprise revenue, insurance companies, healthcare providers, or the business case for safety-first positioning. He keeps the frame exclusively on national security and democratic values, avoiding any suggestion that this is a commercial calculation dressed as principle.

He never mentions specific legal strategies beyond noting he will “challenge it in court” [1]. He does not preview arguments, cite statutes, or name lawyers. Again, this is discipline. The legal fight is coming, and nothing said on camera should constrain it.

And he never discusses the technology itself in technical detail. No model names. No capabilities. No benchmarks. The “basic unpredictability” framing is as far as he goes. He keeps the audience focused on the governance question, not the engineering one.

Analysis: What This Interview Actually Achieved

In my view, this interview was not primarily about explaining Anthropic’s position. It was about establishing a public record for three audiences simultaneously.

For customers and partners, the message was: we are stable, the impact is manageable, we will be fine. Continue doing business with us.

For the courts, the message was: we acted reasonably, we offered continuity, we were given an unreasonable ultimatum, the government’s actions were unprecedented and retaliatory. Every element of a legal challenge was laid out without explicitly making one.

For the public and the technology industry, the message was: there are things we will not build, regardless of who asks. This is what it looks like when a company holds that line.

Whether you find this admirable or calculated — and it is clearly both — the interview is a masterclass in high-stakes corporate positioning. Amodei managed to appear principled without being preachy, defiant without being aggressive, and confident without being dismissive of a genuine threat.

The question that remains is whether the principle survives the pressure. That depends not on what Amodei said in this interview, but on what Anthropic does in the months ahead.

Disclaimer: This article is based on a single-source interview transcript from 3 March 2026. Several claims made during the interview — including Anthropic being the first company on the classified cloud, the three-day ultimatum timeline, and the absence of formal notification — require independent verification. This does not constitute legal, financial, or professional advice.

If your organisation needs support building AI governance frameworks or assessing AI vendor dependency risks, Arkava helps mid-market enterprises turn AI investment into measurable, accountable outcomes.

References

[1] Dario Amodei, CEO of Anthropic. Interview transcript, 3 March 2026. Source: Research Pack provided to The Control Layer.

Why It Matters

The moment a foreign government can unilaterally strip you of access to critical technology — not through technical failure, but through political decree — you don’t really own your AI infrastructure. You’re renting it at the pleasure of another nation’s executive office. On 27 February 2026, the Trump administration demonstrated this principle with surgical precision. And the UK, currently piloting Claude for GOV.UK services, is directly in the line of fire.

Join The Control Layer for weekly perspectives on AI, cybersecurity, and building technology that serves human purpose.

The Move: Federal Ban and Supply-Chain Designation

On 27 February, the Truth Social post landed like a photon torpedo on the AI industry’s hull. The language was blunt – “Leftwing nut jobs,” “Radical Left AI company” – and the directive was absolute: every federal agency was ordered to “immediately cease” using Anthropic technology. The Pentagon got a slightly gentler timeline – six months to complete phase-out – but the message was identical. Stop using Claude. Now.

This wasn’t a quiet policy memo. It came with teeth. And as shown in the post above, it came with the full theatrical force of presidential social media.

Hours later, Defence Secretary Pete Hegseth delivered the formal designation: Anthropic was a “Supply-Chain Risk to National Security.” That label — historically reserved for adversarial foreign entities like Huawei — triggers a nuclear option in military contracting. No US Department of Defense contractor, supplier, or partner may conduct any commercial activity with Anthropic. Full stop. Hegseth also threatened to invoke the Defence Production Act, a Cold War-era statute designed to conscript entire industries for national emergencies.

For context: Anthropic had just become the first frontier AI company to be granted access to US classified networks and National Laboratories. The company had built custom national security models. And it had offered to work directly with the Pentagon on R&D for reliability improvements. All rejected.

Why Anthropic Said No

On 26 February, CEO Dario Amodei published his company’s two red lines.

• First: no mass domestic surveillance.

• Second: no fully autonomous weapons systems.

These aren’t abstract principles. They’re hard limits on what Anthropic will build — and hard limits on what it will do for governments, even America’s.

The company made its position clear: it could have modified its safety guardrails to please the Pentagon. Instead, it chose to face the consequences of saying no.

This is crucial. Because within hours, the market provided a counter-example.

The Pivot: OpenAI and Grok Fill the Void

Sam Altman announced a Pentagon classified network deal for OpenAI on the very day the Anthropic ban dropped. The optics were exquisite – or catastrophic, depending on your view. But Altman made a calculated move: he posted OpenAI’s safety commitments publicly, as shown in his statement above. The company would maintain its prohibition on mass surveillance and autonomous weapons.

Altman wrote that the Department of War “agrees with these principles, reflects them in law and policy, and we put them into our agreement.” He also called on the Pentagon to offer the same terms to all AI companies – a not-so-subtle olive branch to Anthropic.

So why was OpenAI acceptable where Anthropic wasn’t?

The visible difference: OpenAI accepted a closer working relationship, agreed to cloud-only deployment, and offered to station Forward Deployed Engineers at Pentagon facilities. The cynical reading – and in my view, the more accurate one – is that OpenAI was willing to dress up the same red lines in more palatable language whilst maintaining a posture of cooperation rather than confrontation.

But Elon Musk’s xAI made no such rhetorical concessions. Grok signed a Pentagon agreement with a far looser standard: “all lawful use.”

That’s the escape hatch Anthropic refused.

It means whatever the US government deems lawful — today, or tomorrow, or in six months when a new administration changes the definition — is permitted.

The Message to the UK and Europe

Here’s what the UK government should notice: they piloted Claude for a service that guides citizens through government processes. An agentic system. Built on Anthropic’s technology. Anthropic works with the UK AI Security Institute. The company signed a memorandum of understanding with the UK government in February 2025.

As of 27 February 2026, that partnership exists in legal and operational limbo. The US doesn’t directly control UK purchasing decisions. But it controls whether Anthropic can legally serve UK clients — and whether US-based contractors the UK depends on can maintain relationships with Anthropic.

The CLOUD Act means US-headquartered companies are subject to US government orders regardless of where data is hosted. The Data (Use and Access) Act 2025 reinforced UK sovereignty requirements. Those two frameworks are now in direct tension. If a US court orders Anthropic to provide data or restrict access, the company has no good options: comply with America, or break UK law.

The EU faces the same structural problem, amplified. The European Union is >80% dependent on non-European digital infrastructure. The French and German Digital Sovereignty Summit in November 2025 was explicit: Europe is exposed.

Gartner forecasts that by 2027, one-third of enterprises will have moved to localised AI platforms — an explosion from today’s 5%.

And this week, Foreign Policy (A US magazine, founded in 1970, that analyses global affairs, international policy, economics and security) published a piece with a title that says everything: “Europe’s Digital Sovereignty Means Decoupling From U.S. Technology.”

The Trump administration just demonstrated why.

Analysis: Why This Isn’t a Simple Technical Toggle

In my view, the business logic here matters as much as the political logic – and reveals why Anthropic’s stance is rational, not merely idealistic.

Altering Claude’s constitutional AI – the set of principles guiding its behaviour – is not a simple toggle you flip like switching off the Death Star’s exhaust port vulnerability. It’s a significant technical and operational feat. Anthropic would need to retrain models, audit outputs, validate safety behaviours, and manage fallout across thousands of customers globally.

If the Pentagon had demanded Anthropic alter its safety stance specifically for US government use, Anthropic would have faced a choice:

1. accept the demand and incur retraining costs for a single customer, or

2. refuse and lose the entire US government contract.

More likely: Anthropic would have been forced to apply changes globally. It’s simpler — and cheaper — to alter Claude once than to maintain two separate versions of the same model. This means a decision made by the Trump administration could ripple through UK health services, French banks, German manufacturers, and Australian public servants.

This is the sovereignty trap in plain view. Think of it as the Kobayashi Maru of AI governance – a no-win scenario unless you’ve thought about it in advance.

Anthropic’s safety-first positioning has driven real enterprise growth. Insurance companies, healthcare providers, and financial institutions have chosen Claude because its constitutional principles are transparent and non-negotiable. That’s not just ethics. That’s a business differentiator in a market where clients are increasingly worried about being conscripted into surveillance architectures they never signed up for.

But it’s also a liability when the US government decides safety guardrails are inconvenient. Anthropic has built its entire go-to-market strategy around being the “safe AI.” Enterprise customers chose it precisely for the guarantees that the Pentagon now wants removed. Capitulating would likely cost Anthropic more in commercial revenue than the US government contracts are worth.

Risks and Constraints

For UK organisations: The GOV.UK pilot creates a direct dependency on a US-based company during a period of maximum political volatility. If the Trump administration escalates — threatening legal action against UK entities doing business with Anthropic, freezing assets, or pursuing diplomatic pressure — the UK government faces a choice between sovereignty and continuity.

For EU organisations: They face the same problem at scale. Companies with US operations can be squeezed through Commerce Department restrictions or OFAC sanctions. These aren’t hypothetical tools. The Trump administration has signalled willingness to use them.

For Anthropic itself: The company has legal grounds to challenge Hegseth’s designation. The Defence Secretary likely lacks statutory authority to unilaterally bar military contractors from all commercial activity with a US company. But litigation is slow. Market confidence is fast. Customers will hedge by diversifying away from Anthropic.

For the industry: The precedent is now set. Any frontier AI company can be designated a “supply-chain risk” by fiat. The category is newly weaponised.

What to Do Next

For UK Government and Organisations

1. Audit dependencies. Which critical processes depend on US-headquartered AI providers? Map your exposure systematically.

2. Develop redundancy. Don’t move everything to open-source alternatives overnight — that’s unrealistic. But ensure you have fallback options for mission-critical systems.

3. Strengthen data residency. If data is hosted in the UK, can it be legally accessed by US authorities? The answer may be yes. Design around it.

4. Engage in diplomacy. The US-UK relationship is strong enough to merit a conversation about technology infrastructure. Don’t wait for the next crisis.

For EU Organisations

1. Accelerate sovereign alternatives. The French government is investing heavily in European AI models. Support that ecosystem.

2. Assume worst-case scenarios. If a US provider is restricted tomorrow, what happens to your service? Plan for it.

3. Look to partnerships. Japanese, South Korean, and Canadian AI companies offer alternatives with weaker US entanglement.

For Anthropic and Other US Providers

The choice is now starkly visible: maintain safety principles and lose government contracts, or alter them and maintain access. Neither option is good. But the company has chosen the first, and that matters.

Timeline and Implications

This is an evolving situation. The Trump administration has a history of reversing course, negotiating, or softening positions through back channels. Anthropic has signalled it will challenge the Pentagon designation in court. The legal process could take months.

But the market has already moved. Grok and OpenAI have positioned themselves as available alternatives. Customers globally will begin hedging their bets. And governments that depend on US technology will accelerate their sovereignty agendas.

The UK’s GOV.UK pilot serves as a case study in miniature. If it survives this crisis intact, it demonstrates the strength of the US-UK relationship. If it’s disrupted, it becomes a cautionary tale about digital dependency.

Either way, the question is no longer theoretical: who controls your AI?

The answer, it turns out, can be changed by one person’s social media post.

Meanwhile, 573 Google employees and 93 OpenAI employees have signed a “We Will Not Be Divided” open letter supporting Anthropic’s position [14]. The workforce revolt suggests that even companies which accepted Pentagon terms may face internal pressure to maintain the same red lines Anthropic drew publicly.

The AI industry’s conscience, it seems, is distributed across thousands of engineers who’d rather not build Skynet on their lunch break.

Compliance Considerations

Organisations should:

• Review vendor stability clauses and exit provisions in AI service contracts

• Ensure contractual language permits migration to alternative providers without penalty

• Document all AI system functionality to facilitate rapid transition if needed

• Establish internal AI governance structures that don’t depend on external vendors’ policy decisions

Disclaimer: This article represents analysis based on publicly available information as of February 2026. It does not constitute legal, financial, or professional advice. The situation described is evolving rapidly and may have changed since publication.

If your organisation needs support implementing AI governance frameworks or assessing vendor dependency risks, Arkava helps mid-market enterprises turn AI investment into measurable business outcomes.

This is a developing story. The Trump administration has a well-documented history of reversing course on executive actions. The Control Layer will update as new information emerges.

Why it matters: Late last year, Russian state actors deployed wiper malware across 30 Polish wind, solar, and combined heat and power facilities — the first documented destructive cyberattack on distributed renewable energy resources. Operators lost command and control of their systems during an extreme cold period. If your organisation operates, supplies, or depends on distributed energy infrastructure, the defence model you rely on was just proven inadequate.

The attack nobody expected

For a decade, the energy sector’s cybersecurity playbook was written around centralised power stations. Ukraine’s grid attacks in 2015 and 2016 targeted transmission-level infrastructure — large, visible, and (eventually) defensible. Poland’s incident rewrites the threat model entirely.

In late December 2025, attackers attributed to Berserk Bear (Russian FSB) and Sandworm (Russian GRU) compromised 30 distributed energy resource (DER) facilities across Poland. Initial access came through unpatched FortiGate edge devices without multi-factor authentication, using reused credentials. Once inside, they moved laterally through operational technology (OT) networks exploiting devices running default login credentials — the digital equivalent of leaving every door unlocked with the key under the mat [1][2].

The payload was destructive, not espionage. Wiper malware corrupted firmware on remote terminal units (RTUs), deleted human-machine interface (HMI) data, and effectively blinded operators. The systems kept producing power — wind turbines still turned, solar panels still generated — but nobody could see what was happening or intervene if something went wrong [2].

The systems kept producing power. But nobody could see, monitor, or control them. That is a uniquely dangerous kind of failure.

The timing was not coincidental. Poland experienced extreme cold conditions in late December. Had the attack escalated to disrupt heating systems dependent on combined heat and power facilities, the humanitarian consequences could have been severe [1].

Why distributed energy is uniquely vulnerable

The shift to renewables has created what cybersecurity professionals call an “attack surface expansion problem.” Traditional power stations are few in number, physically secured, and staffed by specialists. Distributed energy resources are the opposite: numerous, geographically dispersed, remotely managed, and often operated by organisations without dedicated security teams.

This is the mid-market accessibility gap in action. A 50MW wind farm operator faces the same threat actors as National Grid, but with a fraction of the budget, expertise, and regulatory attention. Default credentials on OT devices from vendors including Hitachi, Mikronika, and Moxa were exploited in Poland — not because the technology was fundamentally flawed, but because smaller operators lacked the resources and processes to change them [2].

CISA’s emergency response alert, issued 9 February 2026, explicitly flagged DERs as a “novel target category” requiring urgent attention. The alert directed federal agencies and critical infrastructure operators to inventory internet-exposed edge devices, enforce MFA on network management interfaces, and mandate vendor enforcement of password changes during integration [1].

What this means for the UK

The UK’s net zero strategy depends heavily on distributed renewable infrastructure. Offshore and onshore wind, solar farms, battery storage, and smart grid systems are expanding rapidly. Each new facility extends the attack surface — and many are operated by mid-market companies without enterprise-grade cybersecurity programmes.

Three UK-specific risks demand attention.

First, the Cyber Security and Resilience Bill currently progressing through Parliament will, for the first time, bring managed service providers (MSPs) and data centres into scope for mandatory incident reporting. Initial notification within 24 hours of awareness, full report within 72 hours. Fines up to £17 million or 4% of worldwide turnover [5][6]. Energy operators relying on third-party MSPs for remote monitoring of DER facilities must ensure their contractual arrangements reflect these incoming obligations — before enforcement begins.

Second, the NCSC’s Cyber Assessment Framework (CAF) v4.0, updated August 2025, establishes requirements for a “definitive record” of OT assets — continuous, trustworthy, and segmented by consequence and exposure. The Poland attack succeeded precisely because asset visibility was inadequate. Operators did not know what devices were internet-exposed, what credentials they were running, or what firmware versions were installed [8].

Third, the convergence of IT and OT in renewable energy creates interdependencies that traditional segmentation models were not designed to handle. Building Information Modelling (BIM), digital twins, IoT sensors, and remote monitoring platforms create what researchers describe as “very high risk” network-of-systems architectures where a single compromise can cascade across entire project portfolios [17].

Analysis

In my view, the Poland attack represents a strategic inflection point rather than an isolated incident. Russian state actors have demonstrated they can — and will — target the distributed infrastructure that underpins modern energy systems. The tactical shift from centralised grid attacks to dispersed, under-resourced facilities is logical from an adversary’s perspective: lower defences, higher volume of targets, and a defender community that has not yet adapted.

The UK energy sector should treat this as a warning, not a headline. The organisations most at risk are not the large utilities with dedicated security operations centres. They are the mid-market operators, independent power producers, and community energy schemes that form the distributed backbone of our renewable infrastructure. These organisations need access to enterprise-grade security capabilities at mid-market price points — which is precisely where the current market fails them.

The regulatory environment is responding. The CSRB’s expanded scope, the NCSC’s updated CAF, and CISA’s emergency directives all point in the same direction: distributed energy security is no longer optional. But regulation alone does not secure systems. Implementation does.

Risks and constraints

Several important caveats apply. Attribution to Berserk Bear and Sandworm carries confirmed confidence from Polish government statements, but independent verification of specific TTPs remains ongoing. The full scope of the compromise may not yet be public — forensic analysis of 30 facilities takes considerable time.

The UK’s regulatory landscape is evolving. The CSRB’s secondary legislation (which will contain the detailed requirements for supply chain risk management) has not yet been published. Organisations should prepare for compliance but should not assume final requirements will mirror current drafts exactly.

Finally, the comparison between Polish and UK energy infrastructure has limits. Grid architectures, regulatory frameworks, and operational practices differ. Poland’s experience is instructive, not directly predictive. The principle — that distributed systems require distributed defence — is universal. The implementation details will be UK-specific.

What to do next

For boards and executives: Commission an immediate assessment of your organisation’s exposure to distributed OT infrastructure risk. Ask three questions: Do we know every internet-connected device in our operational technology estate? Do any of them run default credentials? What is our contractual position with MSPs regarding incident notification under the incoming CSRB? If you cannot answer all three with confidence, you have work to do.

For technical leaders: Inventory all internet-exposed edge devices (FortiGate, Cisco, SonicWall appliances) across your OT environment. Enforce MFA on every network management interface. Change every default credential — not next quarter, now. Implement firmware verification on OT devices and segment your network DMZ from internal environments. CISA’s emergency directive provides a practical checklist [1].

For mid-market energy operators: You face the same threat actors as national utilities with a fraction of their resources.

Prioritise three actions:

1. Asset visibility — you cannot defend what you cannot see;

2. Edge device hardening — unpatched, internet-exposed devices with default credentials are the primary attack vector;

3. MSP contract review — ensure your suppliers’ incident response obligations align with 24-hour CSRB notification requirements before enforcement begins.

Disclaimer: This article represents analysis based on publicly available information as of February 2026. It does not constitute legal, financial, or professional advice.

If your organisation needs support implementing OT security frameworks or preparing for CSRB compliance, Arkava helps mid-market enterprises turn cybersecurity investment into measurable resilience outcomes.

References

[1] CISA. “OT Vulnerabilities Alert — Poland Energy Grid Incident.” CISA, 9 February 2026.

[2] Industrial Cyber. “CISA Alerts on OT Vulnerabilities After Poland Energy Attack Damaged RTUs and Wiped HMI Data.” Industrial Cyber, 10 February 2026.

[5] Osborne Clarke. “UK Cyber Security and Resilience Bill: Pragmatic Overhaul or Regulatory Overload?” Osborne Clarke, 10 December 2025.

[6] OpenPolicy. “UK Cybersecurity Shake-Up: What a New Cyber Security & Resilience Bill Signals for 2026.” OpenPolicy, 19 November 2025.

[8] NCSC. “Annual Review 2025.” UK National Cyber Security Centre, October 2025.

[17] ScienceDirect. “A Perfect Storm: Digital Twins, Cybersecurity, and General Contracting Firms.” Developments in the Built Environment, 2024.

Why it matters: Accenture's acquisition of Faculty – the UK's most prominent AI-native firm – is not an isolated deal. It is the twenty-third AI acquisition Accenture has completed in 2025 alone, part of a broader consolidation wave that includes Microsoft's $250 billion restructuring of its OpenAI relationship. For mid-market organisations still experimenting with AI, the message is stark: the builders are being bought, the infrastructure is being locked in, and the window to establish independent AI capability is narrowing fast.

Join The Control Layer for weekly perspectives on AI, cybersecurity, and building technology that serves human purpose.

The acquirers are buying the builders

Faculty was, until last week, the closest thing the UK had to a sovereign AI consultancy with genuine government-grade credentials. Founded in 2014, it built its reputation on applied AI for the UK public sector, defence, and regulated industries. Its work with the NHS during COVID-19 and subsequent defence contracts established it as the go-to firm when UK organisations needed AI capability they could trust [1].

Accenture paid approximately £500 million for that trust. The acquisition follows a pattern that should concern every mid-market technology leader in the country: the large consultancies and hyperscalers are systematically absorbing the independent AI talent pool [2].

This is not speculation. Accenture alone has completed twenty-three AI-related acquisitions in 2025, building what amounts to a vertically integrated AI services empire [2]. Meanwhile, Microsoft’s $250 billion restructuring of its OpenAI partnership signals that even the largest technology relationships are being reconfigured around AI infrastructure ownership [3].

The pattern is unmistakable: enterprise AI is moving from the experimentation phase to the infrastructure phase, and the infrastructure is being consolidated into fewer, larger hands.

What consolidation actually means for mid-market organisations

When a firm like Faculty gets acquired, the immediate press coverage focuses on the deal value and the strategic rationale for the acquirer. What rarely gets discussed is the downstream effect on the organisations Faculty used to serve directly.

Three things happen when independent AI capability gets absorbed into a global consultancy.

First, pricing changes. Faculty’s mid-market clients could access genuine AI expertise at rates that reflected a 200-person specialist firm. Those same capabilities, delivered through Accenture’s engagement model, come with enterprise pricing, minimum contract sizes, and overhead structures designed for FTSE 100 clients, not £50 million revenue businesses.

Second, attention shifts. Global consultancies serve global clients. The bespoke attention that a mid-market UK manufacturer received from Faculty’s team will, inevitably, compete for resources against larger, more profitable engagements. This is not a criticism of Accenture’s intentions – it is a structural reality of how large professional services firms allocate talent.

Third, and most significantly, dependency deepens. When the independent options disappear, mid-market organisations have fewer choices. Build capability internally, pay enterprise rates to the consolidators, or do nothing and fall further behind.

The 75% failure rate makes this worse, not better

Here is where consolidation intersects with a problem the industry has been quietly ignoring. Industry benchmarks consistently show that 75% of AI investments fail to deliver measurable business value [4]. That statistic predates the current consolidation wave. It describes a market where organisations had more choice, more independent advisers, and more competitive pressure on implementation quality.

Consolidation does not inherently solve the value delivery problem. In some cases, it exacerbates it. When fewer providers control more of the market, the competitive incentive to tie fees to outcomes diminishes. Why guarantee results when clients have limited alternatives?

The architecture, engineering, and construction sector provides a useful case study. This week’s data shows 74% AI adoption across AEC firms, a figure that sounds impressive until you examine what “adoption” means in practice. For many organisations, adoption means purchasing licences and running pilot projects. It does not mean delivering measurable operational improvement. The gap between AI adoption and AI value is where most organisations lose money, and consolidation makes that gap harder to close.

Analysis

In my view, this consolidation wave represents the most significant structural shift in the UK AI services market since the original cloud migration era of 2015-2019.
The parallel is instructive. During cloud migration, mid-market organisations that moved early and built internal competence ended up with genuine flexibility and competitive advantage. Those that waited and relied entirely on external providers found themselves locked into expensive, inflexible arrangements that took years to unwind.

AI capability is following the same trajectory, but faster. The firms that could have helped mid-market organisations build independent AI competence are being absorbed into structures designed to serve enterprise clients. The talent pool is not growing fast enough to replace them. And the technology itself is advancing at a pace that makes delayed action increasingly expensive.

The simultaneous launch of Claude Opus 4.6’s multi-agent orchestration, OpenAI’s Frontier enterprise service, and Mistral’s on-device models in a single week illustrates the acceleration. These are not incremental updates. They represent fundamental shifts in what AI systems can do autonomously. Organisations that lack the internal capability to evaluate, implement, and govern these tools will find themselves dependent on whoever can – and after this consolidation wave, the “whoever” is an increasingly small group charging increasingly large fees.

The EU’s decision to delay high-risk AI compliance deadlines [5] adds another dimension. Regulatory uncertainty benefits large organisations with dedicated compliance teams. It penalises mid-market firms that were counting on clear regulatory timelines to plan their AI governance investments. Every delay extends the period of ambiguity, and ambiguity favours those with the deepest pockets.

Risks and constraints

The consolidation narrative has legitimate counter-arguments. Large consultancies bring scale, methodology, and risk management capabilities that smaller firms cannot match. Accenture’s acquisition of Faculty may genuinely improve the quality and reliability of AI delivery for some clients, particularly in defence and high-security environments where Accenture’s global compliance infrastructure adds value.

There is also a talent argument. Faculty’s engineers may have more career development opportunities within a global firm, which could improve retention and attract stronger candidates. Better talent, in theory, means better outcomes for clients.

However, these benefits accrue primarily to large enterprise clients. For mid-market organisations, the risk profile is weighted differently. The critical constraint is time: the window to build internal AI capability while independent advisory options still exist is measurably shorter after this week than it was before.

It is also worth noting that Faculty’s government contracts introduce sovereignty considerations. UK defence and public sector AI capability is now partially owned by a US-headquartered consultancy. Whether this matters depends on one’s view of sovereign technology requirements, but it is a factor that boards in regulated sectors should consider.

What to do next

For boards and executives: Commission an honest assessment of your organisation’s AI capability dependency. Map every external AI provider relationship and ask: if our primary provider doubled their rates or shifted focus to larger clients, what would we do? If the answer is “nothing, because we have no alternatives,” that is your strategic risk. This quarter, not next year.

For technical leaders: Evaluate the build-versus-buy equation with fresh assumptions. The cost of building internal AI capability has decreased significantly with open-source models and platforms. The cost of external dependency has increased and will continue to increase as consolidation reduces competition. Run the numbers again – the answer may have changed.

For mid-market organisations: Consider whether outcome-based AI partnerships – where fees are tied to measurable business results rather than time and materials – offer a structural solution to the consolidation problem. Providers willing to guarantee outcomes have a fundamentally different incentive structure from those selling capacity. In a consolidating market, aligned incentives become your most important procurement criterion.

Disclaimer: This article represents analysis based on publicly available information as of February 2026. It does not constitute legal, financial, or professional advice.

If your organisation needs support building independent AI capability before the consolidation window closes, Arkava helps mid-market enterprises turn AI investment into measurable business outcomes – with pricing tied to results, not billable hours.

References

[1] Faculty AI. Company background and UK public sector portfolio. faculty.ai, 2026. [2] Accenture. AI acquisition programme; Faculty acquisition announcement, February 2026.
[3] Microsoft-OpenAI. Partnership restructuring and $250 billion investment framework, February 2026.
[4] Industry benchmark data. AI investment value realisation rates across enterprise deployments.
[5] European Commission. EU AI Act high-risk compliance timeline adjustment, February 2026.

Why it matters: The NCSC handled 204 nationally significant cyber incidents in twelve months – a 130% increase on the previous year. The Cyber Security and Resilience Bill is now in committee stage, expanding obligations to data centres, managed service providers, and critical suppliers. If your organisation falls into any of those categories, the compliance clock has started.

— Get The Control Layer in your inbox — weekly insights on AI, cybersecurity, and digital transformation: TheControlLayer.arkava.ai

The numbers nobody can ignore

The NCSC Annual Review, published in October 2025, delivered the most candid official assessment of UK cyber readiness in a decade. The figures tell a story that even the most optimistic boardroom cannot dismiss: 204 nationally significant incidents between September 2024 and August 2025, compared with 89 the previous year [1]. Eighteen of those were classified as highly significant – a 50% increase from twelve the year before [1].

To put this in practical terms, the UK experienced an average of four nationally significant cyber attacks every week. Not minor probing. Not unsuccessful attempts caught at the perimeter. Incidents serious enough to warrant national-level coordination.

NCSC CEO Richard Horne described the situation with unusual directness, stating that cybersecurity is now a matter of business survival and national resilience [1]. GCHQ Director Anne Keast-Butler reinforced this by referencing the attacks on M&S, Co-op, and Jaguar Land Rover as a stark reminder of what happens when defences fail [1].

The 204-incident figure represents a 130% year-on-year increase in nationally significant cyber incidents. This is not a statistical anomaly – it reflects a structural escalation in threat capability and targeting.

Who is behind it, and how they operate

The NCSC assessment identifies China as the biggest state-level cyber threat to the United Kingdom [1]. However, the most immediate operational risk for mid-market organisations comes not from state actors directly, but from the ransomware ecosystem they have helped to professionalise.

The DragonForce group’s attacks on UK retail businesses demonstrated a pattern now familiar across threat intelligence reports: sophisticated initial access, lateral movement through poorly segmented networks, and double-extortion tactics combining data theft with encryption [1]. The rise of malware-as-a-service platforms has democratised access to these capabilities, meaning that threat actors with limited technical expertise can now deploy tools previously reserved for well-resourced groups [1].

For boards and security leaders, the implication is clear. The question is no longer whether your organisation is a target – at mid-market scale, you almost certainly are.
The question is whether your defences reflect the reality of 2026 threat operations or rather the assumptions of 2020.

The legislative response: Cyber Security and Resilience Bill

The government’s response to this escalation is the Cyber Security and Resilience (Network and Information Systems) Bill, introduced on 12 November 2025 [2]. After its Second Reading on 6 January 2026, the Bill entered Public Bill Committee on 3 February 2026, with the committee expected to report by 5 March 2026 [2]. Royal Assent is anticipated by mid-2026.

This is not incremental policy refinement. The Bill significantly expands the scope of the existing NIS Regulations 2018 to include data centres, medium and large managed service providers, critical suppliers, and large load controllers [2]. If your organisation provides managed IT services to regulated sectors, or operates data centre infrastructure, you are moving from voluntary best practice to statutory obligation.

The Bill was announced in the King’s Speech 2024 as part of the government’s Plan for Change, signalling that cyber resilience is now treated as critical national infrastructure policy rather than a technology niche [2].

£210 million and a change in approach

Alongside the Bill, the government published its Cyber Action Plan on 6 January 2026, committing £210 million to a new Government Cyber Unit that will centralise cross-government cyber risk management [3]. The plan introduces a Software Security Ambassador Scheme and a Software Security Code of Practice designed to address supply chain attacks – the vector responsible for some of the most damaging incidents in the NCSC’s review period [3].

The Cyber Profession initiative aims to tackle the talent pipeline problem that mid-market organisations know intimately. When you cannot compete with City salaries or government security clearance pathways, retaining cyber professionals becomes an existential challenge for organisations in the £50M–£500M revenue band.

Analysis

Here is what concerns me about these developments, and it is not the headline number.

The 204 incidents occurred during a period when most organisations believed they were improving their security posture. Spending on cyber tools has increased year on year across UK business. Boardroom awareness has never been higher. And yet the incident curve moved sharply upward.

This suggests a structural mismatch between how organisations invest in cyber resilience and how threat actors actually operate. The investment goes into tools. The attacks exploit people, processes, and architectural decisions made five years ago. You can buy the most sophisticated endpoint detection platform available and still fall to a compromised managed service provider in your supply chain – which is precisely the gap the new Bill attempts to close.

The timing of the Bill’s committee stage is significant. Organisations that wait for Royal Assent to begin compliance preparation will find themselves scrambling. Those that start now – mapping their supply chain exposure, auditing managed service provider contracts, assessing data centre dependencies – will be better positioned both for compliance and for genuine resilience.

Risks and Constraints

The 130% increase in nationally significant incidents deserves context. The NCSC also handled 429 total incidents, virtually unchanged from 430 the previous year [1]. This suggests the severity of incidents escalated significantly while overall volume remained stable. The increase in nationally significant classification may partly reflect changes in triage methodology and reporting thresholds, not solely an increase in attack volume.

The Cyber Security and Resilience Bill is still in committee. Amendments may alter scope, timelines, and enforcement mechanisms before Royal Assent. Organisations should prepare for the direction of travel without over-investing in specific compliance interpretations that may shift.

The £210M Cyber Action Plan investment, while substantial, is distributed across government. Its impact on private-sector mid-market organisations will be indirect, primarily through improved threat intelligence sharing and supply chain standards.

What to Do Next

For boards and executives: Request a board-level briefing on the NCSC Annual Review findings and the Cyber Security and Resilience Bill’s scope. Specifically ask whether your organisation or any of your critical suppliers fall within the expanded NIS regulations. This is a governance question, not a technology question.

For technical leaders: Map your managed service provider and data centre dependencies against the Bill’s expanded scope. Begin supply chain risk assessments now – contractual renegotiations take months, and you will want updated terms before enforcement begins.

For mid-market organisations: The window between now and Royal Assent (approximately mid-2026) is your preparation period. Commission an independent assessment of your cyber resilience posture against the new requirements. The organisations that treat this period as a head start rather than a waiting period will spend less, disrupt operations less, and comply faster.

Disclaimer: This article represents analysis based on publicly available information as of February 2026. It does not constitute legal, financial, or professional advice.

If your organisation needs support assessing its exposure to the Cyber Security and Resilience Bill’s expanded requirements, Arkava helps mid-market enterprises turn cyber resilience investment into measurable business outcomes.

References

[1] NCSC. “NCSC Annual Review 2025.” National Cyber Security Centre, 14 October 2025.
[2] UK Parliament. “Cyber Security and Resilience Bill.” Parliament.uk, 2025-2026.
[3] DSIT. “Cyber Action Plan.” Department for Science, Innovation and Technology, 6 January 2026.

Why it matters: Ransomware gangs have industrialised insider recruitment, offering your IT staff 15–25% of ransom payments for administrative access. This transforms insider threat from theoretical edge case to primary attack vector — and most organisations' governance frameworks are unprepared.

Join The Control Layer for weekly perspectives on AI, cybersecurity, and building technology that serves human purpose.

The job offer your security engineer won’t mention

Somewhere in your organisation, someone with privileged access has received an interesting proposition. Not through a dark web forum or encrypted chat — but through LinkedIn, professional networks, or even direct email to their corporate inbox.

The pitch is straightforward: provide credentials, disable backup systems, or simply leave a door open. In return, receive a percentage of whatever ransom your employer eventually pays. Anonymous payment in cryptocurrency. No face-to-face meetings. Minimal risk of detection.

This is not hypothetical. In December 2025, BlackCat/ALPHV successfully recruited a cybersecurity professional — someone whose literal job was preventing exactly this kind of attack — to facilitate a breach. The individual received a reported 20% of the ransom payment.[1]

Welcome to the industrialisation of insider threat.

From opportunism to operating model

For years, insider threat occupied a peculiar position in enterprise risk registers: acknowledged as dangerous, yet treated as unpredictable. You might screen employees at hiring, monitor for anomalous behaviour, and hope that loyalty or fear of consequences would deter betrayal.

Ransomware-as-a-service gangs have rendered this approach obsolete.

NCC Group’s January 2026 analysis reveals that major RaaS operations — Qilin, Medusa, and the remnants of BlackCat/ALPHV — now run formalised recruitment programmes.[1] These are not ad-hoc approaches to disgruntled employees. They are structured affiliate expansion efforts with defined compensation models, anonymity guarantees, and target profiles.

The economics are compelling. A ransomware gang investing in zero-day exploitation, phishing infrastructure, and technical penetration might spend weeks or months establishing access to a target environment. An insider with administrative credentials can provide equivalent access in an afternoon. The cost-benefit calculation has shifted decisively toward human exploitation.

The uncomfortable truth: your perimeter defences assume the threat is outside. Insider recruitment programmes target people who are already inside your most sensitive systems.

Who they’re targeting — and why it matters for governance

The recruitment focus is disturbingly strategic. RaaS gangs are not approaching random employees; they are targeting specific roles with maximum leverage:

IT administrators and security engineers represent the primary target class. These individuals possess the technical knowledge to disable monitoring, create persistent access, and — critically — understand which actions will trigger alerts and which will not. The December 2025 BlackCat case demonstrates that even professionals whose identity is built around defending organisations can be compromised.[1]

HR and identity management staff control credential issuance, onboarding processes, and employee records. Access to these systems enables creation of ghost accounts, manipulation of access permissions, and — in more sophisticated scenarios — establishment of entirely fictitious employees with legitimate system access.

Backup operators and disaster recovery personnel hold the keys to organisational resilience. Ransomware is only effective if victims cannot restore from backups. An insider who degrades backup integrity before an attack transforms a recoverable incident into an existential crisis.

The pattern suggests reconnaissance. Threat actors are not casting wide nets; they are identifying specific individuals with specific capabilities, then making tailored approaches. This implies pre-attack research into organisational structures, likely through OSINT gathering from LinkedIn, corporate websites, and professional publications.

The compensation model: why your salary might not be enough

The financial incentives are substantial and deliberately calibrated. NCC Group documents offers ranging from 15–25% of eventual ransom payments, with cryptocurrency settlement and promised anonymity.[1]

Consider the arithmetic. A mid-market organisation facing a £2 million ransom demand — not unusual for companies in the £50–500 million revenue band — represents a potential £300,000–£500,000 payout for a cooperating insider. For an IT administrator earning £60,000 annually, this represents five to eight years of salary for a single act of betrayal.

The moral calculus is further complicated by rationalisation opportunities. Ransomware gangs explicitly position their recruitment as victimless: “The company has insurance. No one gets hurt. You’re just accelerating an inevitable outcome.” These narratives provide psychological cover for individuals who might otherwise resist.

This is not to suggest your teams are inherently untrustworthy. The vast majority of employees would never consider such offers. But governance cannot rely on universal virtue. It must assume that among any sufficiently large workforce, some individuals will face financial pressure, personal grievances, or moral flexibility that makes recruitment approaches viable.

Analysis: why traditional insider threat programmes fail

Most organisations’ insider threat controls were designed for a different problem — the disgruntled employee who acts alone, or the careless user who accidentally exposes data. These controls assume insider threats emerge from internal conditions: workplace conflict, personal crisis, or ideological motivation.

The RaaS recruitment model inverts this assumption. The threat originates externally, targeting individuals who may have no prior indicators of risk. Traditional behavioural monitoring looks for anomalies against baseline behaviour; a well-coached insider can maintain normal patterns until the moment of betrayal.

Three structural weaknesses emerge:

Privilege accumulation without accountability. Many organisations allow administrative credentials to persist indefinitely, with limited logging of privileged actions. An insider can conduct preparatory activities — disabling alerts, mapping backup systems, creating persistence mechanisms — over weeks or months without triggering review.

Backup systems as single points of failure. The same individuals responsible for backup operations often possess the access to compromise those backups. When the attacker is the backup administrator, the assumption that “we can always restore” becomes dangerously optimistic.

Trust-based access models. Enterprise security architectures frequently assume that authenticated users with appropriate role assignments should receive broad access within their functional domain. This model collapses when the authenticated user is cooperating with external threat actors.

Risks and constraints

Addressing insider recruitment requires acknowledging uncomfortable organisational realities.

Zero trust creates friction. Implementing genuine least-privilege access, continuous authentication, and comprehensive logging of privileged actions will slow operations. Users will complain. Productivity metrics may decline. Leadership must accept these trade-offs as the cost of resilience.

Monitoring has limits. Even sophisticated user behaviour analytics cannot reliably detect a well-coached insider acting within their authorised access patterns. Technical controls are necessary but insufficient; governance frameworks must assume some insider compromises will occur and design for containment.

Cultural implications are significant. Communicating to employees that the organisation is implementing enhanced monitoring because “we cannot fully trust anyone” requires careful handling. Poorly executed messaging can damage morale and, ironically, increase insider risk by eroding organisational loyalty.

Resource constraints are real. Mid-market organisations lack the security operations centre staffing to implement comprehensive privileged access monitoring. Solutions must be proportionate to organisational capacity — which often means accepting residual risk and focusing controls on highest-impact assets.

What to do next

The shift from opportunistic insider threat to industrialised recruitment demands corresponding evolution in governance and controls.

For boards and executives: Commission an independent assessment of privileged access governance. Ask specifically: who has administrative access to backup systems, and what oversight exists for their actions? Ensure cyber risk registers explicitly address insider recruitment as a threat vector, not merely “insider threat” as an abstract category. Consider whether director-level accountability for insider threat governance is adequately defined.

For security leaders: Implement privileged access management with just-in-time elevation and comprehensive session recording. Establish separation of duties for backup operations — no single individual should be able to both manage backups and delete backup logs. Deploy user and entity behaviour analytics focused on privileged account classes, accepting that detection rates will be imperfect. Conduct tabletop exercises specifically modelling insider-facilitated ransomware scenarios.

For mid-market organisations with limited resources: Focus controls on your crown jewels. Identify the three to five systems whose compromise would be existential, and implement disproportionate monitoring and access controls for those specific assets. Consider managed detection and response services that provide 24/7 privileged access monitoring without requiring internal SOC staffing. Ensure backup systems include immutable storage that cannot be modified even by backup administrators.

For HR and people functions: Review pre-employment screening for roles with privileged access, recognising that past behaviour is an imperfect predictor of future compromise. Establish confidential reporting channels for employees who receive suspicious approaches. Consider whether exit procedures adequately address credential revocation and knowledge transfer risks.

Disclaimer: This article represents analysis based on publicly available information as of January 2026. It does not constitute legal, financial, or professional advice.

If your organisation needs support implementing insider threat governance frameworks, Arkava helps mid-market enterprises turn security investment into measurable risk reduction.

References

[1] NCC Group. “Ransomware Report: December 2025.” January 2026.

[2] National Cyber Security Centre (NCSC). “NCSC Annual Review 2025.” October 2025.

[3] Cybersecurity and Infrastructure Security Agency (CISA) and Federal Bureau of Investigation (FBI). “Joint Cybersecurity Advisory: Ransomware Trends.” December 2025.

[4] The Control Layer (Amer Altaf). “The Zero-Day Supply Chain: What Cl0p’s Oracle Campaign Means for You.” 8 December 2025.

[5] S-RM Cyber Intelligence. “ShinyHunters Launch Vishing Attacks on SSO Accounts.” 30 January 2026.

[6] Acumency Cyber. “Cyber Threat Intelligence Digest – January 2026: Week 4.” 31 December 2025.

Why it matters: Across manufacturing, retail, healthcare, and government, organisations achieving 12-35% ROI share one trait: they prioritised real-time visibility over comprehensive technology replacement. The implication for UK mid-market leaders is clear—your next transformation investment should measure before it builds.

Join The Control Layer for weekly perspectives on AI, cybersecurity, and building technology that serves human purpose.

The pattern hiding in plain sight

Something counterintuitive emerged from the past week’s digital transformation data. The organisations delivering measurable returns—30% reduction in unplanned downtime at VINOVALI, 63% diagnostic improvement in UK healthcare, 13 million users on GOV.UK One Login—didn’t deploy the most sophisticated technology. They deployed the most visible.[1][2][3]

This isn’t semantic wordplay. It’s a strategic distinction with budget implications.

Manufacturing execution systems that delivered 35% ROI within 90 days didn’t automate decision-making. They automated measurement. Retail automation that freed staff for high-value work didn’t replace humans. It offloaded transaction tracking so humans could focus on customers. Government platforms that scaled to millions didn’t build monolithic systems. They built componentised visibility layers—identity, access, AI-assisted navigation—each measurable independently.[1][4][3]

The principle is simple: you cannot optimise what you cannot see. And you cannot see what you’re not measuring.

What the data actually shows

Manufacturing: From guesswork to discipline

VINOVALI, a French food manufacturer, achieved a 30% reduction in unplanned downtime within six months of deploying a cloud-based manufacturing execution system. The technology wasn’t revolutionary. The discipline was.[1]

Before deployment, production and maintenance teams operated on assumption. Losses were considered structural—”that machine always runs hot”—rather than addressable. The MES didn’t introduce new equipment. It introduced loss tracking. Real-time. Automated. Unambiguous.

Once loss patterns became visible, maintenance teams shifted from reactive troubleshooting to predictive intervention. The 30% downtime reduction emerged not from capital investment but from what one implementation partner calls “visibility discipline”—the organisational habit of acting on measured loss rather than assumed loss.[1]

Across hundreds of UK manufacturing deployments, a consistent pattern emerged: 15% operational efficiency gains and two hours daily saved per operator. Not from automation. From eliminating manual spreadsheet reconciliation and decision delays inherent to batch-based reporting.[1][2]

Healthcare: Augmentation over automation

A UK healthcare provider achieved 63% improvement in diagnosis rates through machine learning-based patient monitoring. The instinct might be to credit the algorithm. The reality is more nuanced.[5]

Effective healthcare AI doesn’t automate diagnosis. It augments clinician capability. The 63% improvement reflects ML pattern recognition combined with human clinical judgment—reducing both false positives (purely algorithmic) and false negatives (purely human). The system made patterns visible. Clinicians made decisions.

This distinction matters for governance. Systems designed to replace human judgment face regulatory scrutiny, clinical resistance, and liability questions. Systems designed to make patterns visible face adoption curves.

Retail: Labour redirection, not labour replacement

UK retailers deploying automation reported benefits within 6-12 months, but the metrics that mattered weren’t pure cost reduction. They were operational resilience and customer loyalty.[4][6]

Retailers facing structural labour shortages shifted staff from transactional roles—checkout, basic inventory—to specialist activities: customer consultation, loss prevention, brand engagement. The automation handled routine measurement and tracking. Humans handled exceptions and relationships.

The business outcome: 53% of customers say they’d avoid retailers who fail on order accuracy. Automation that enables accuracy while freeing humans for relationship-building isn’t efficiency theatre. It’s competitive positioning.[4]

Government: Components over monoliths

The UK Government Digital Service published its Roadmap for Modern Digital Government on 26 January 2026, revealing the architecture behind one of the world’s most mature citizen-facing digital platforms.[3][7]

GOV.UK didn’t build a monolithic system. It built visible, measurable components:

• GOV.UK One Login: 13 million users accessing 120+ government services through a single authentication layer

• GOV.UK App: 316,000 downloads by December 2025, providing unified mobile access

• GOV.UK Chat: AI-assisted navigation rolling out across all GOV.UK properties in 2026

Each component is independently deployable, independently measurable, and independently improvable. The strategy prioritises reduced friction over feature maximisation. Citizens experience fewer logins, simpler navigation, and 24/7 AI support—not because of technological sophistication but because of architectural visibility.[3][7]

Analysis: The visibility-first framework

Three implementation patterns distinguish organisations achieving measurable ROI:

Pattern 1: Measurement precedes automation. Every successful case study began by making losses, inefficiencies, or friction visible before attempting to address them. Manufacturing MES tracked losses before optimising production. Retail automation measured transaction patterns before reallocating labour. Government platforms measured user journeys before simplifying authentication.

Pattern 2: Human judgment remains central. Systems that preserved human decision-making while automating measurement achieved faster adoption and higher outcomes than systems designed to automate decisions. Healthcare AI augments clinicians. Retail automation frees staff for customer engagement. Government Chat assists navigation rather than replacing it.

Pattern 3: Phased implementations outperform big-bang deployments. Manufacturers starting with single-line MES pilots delivered 12% ROI within 30 days, 25% by day 60, and 35% by day 90. Retailers deploying store automation in phases reported benefits within 6-12 months. Government platforms scaled through componentised architecture. Big-bang approaches faced 45% adoption rates versus 85% for phased rollouts.[1][2]

Risks and constraints

The visibility principle isn’t a silver bullet.

Data quality remains the bottleneck. Visibility into bad data produces confident wrong decisions. Manufacturing MES implementations assume sensor accuracy. Healthcare AI assumes clean patient records. Government platforms assume identity verification integrity. Each assumption carries risk.

Cultural change is harder than technology change. VINOVALI’s 30% downtime reduction required production and maintenance teams to align around data-driven decisions rather than assumed losses. Not every organisation has the leadership alignment to make that shift.

Vendor lock-in risks compound with componentisation. GOV.UK’s componentised architecture is elegant, but each component creates a dependency. Organisations adopting similar approaches must consider exit strategies alongside entry points.

Compliance creates new visibility requirements. NIS2, the Cyber Security and Resilience Bill, and the EU AI Act all mandate visibility into systems, data flows, and decision processes. Visibility-first approaches align with these requirements, but compliance itself consumes resources.

What to do next

For boards and executives: Challenge any transformation proposal that can’t articulate what will become visible in the first 90 days. If the business case depends entirely on benefits realised after 18+ months, the risk profile is wrong. Phased implementations with early visibility milestones de-risk transformation investments and provide decision points for scaling or pivoting.

For technical leaders: Audit your current transformation initiatives against the visibility principle. Are you measuring before optimising? Are you preserving human judgment in decision flows? Are you deploying in phases that allow independent measurement? If not, consider restructuring.

For mid-market organisations: The visibility principle is your competitive advantage. You can’t outspend enterprise competitors on technology, but you can out-measure them on outcomes. Start with single-line pilots. Demand ROI measurement at 30, 60, and 90 days. Scale what works. Abandon what doesn’t. This is how 15% efficiency gains and two-hour daily savings per operator become your baseline, not your aspiration.[1]

Disclaimer: This article represents analysis based on publicly available information as of January 2026. It does not constitute legal, financial, or professional advice.

If your organisation needs support implementing visibility-first transformation frameworks, Arkava helps mid-market enterprises turn AI investment into measurable business outcomes.

References

[1] ITI Group. “Why Manufacturing Execution Systems should be first on an SME’s digital roadmap in 2026.” ITI Group, 21 January 2026. https://www.itigroup.com/why-manufacturing-execution-systems-should-be-first-on-an-smes-digital-roadmap-in-2026/

[2] Shoplogix. “Phased MES Implementation Approach: Your Step-by-Step Guide.” Shoplogix, 18 September 2025. https://shoplogix.com/phased-mes-implementation-approach/

[3] UK Government Digital Service. “Roadmap for Modern Digital Government.” Department for Science, Innovation and Technology, 26 January 2026. https://www.gov.uk/government/organisations/government-digital-service

[4] Retail Tech Innovation Hub. “New retail efficiency for 2026: transforming operations with AI automations.” Retail Tech Innovation Hub, 15 January 2026. https://retailtechinnovationhub.com/home/2025/12/16/new-retail-efficiency-for-2026-transforming-operations-with-ai-automations

[5] Rishabh Soft. “Top Digital Transformation Trends and Strategies for 2026.” Rishabh Soft, 7 December 2025. https://www.rishabhsoft.com/blog/digital-transformation-trends

[6] Conversations on Retail. “Why 2026 Is Emerging as the Year In-Store Automation Becomes Retail Infrastructure.” Conversations on Retail, 29 January 2026. https://conversationsonretail.com/why-2026-is-emerging-as-the-year-in-store-automation-becomes-retail-infrastructure/

[7] Government Digital Service Blog. “Our roadmap for modern digital government.” GDS Blog, 20 January 2026. https://gds.blog.gov.uk/2026/01/20/our-roadmap-for-modern-digital-government/

Why it matters: This week, the US and Taiwan announced a $250 billion semiconductor investment pact. Singapore committed S$1 billion to AI research through 2030. And Nvidia's planned $100 billion investment in OpenAI stalled amid internal doubts. These are not commercial transactions. They are acts of industrial policy that will determine which nations control the substrate of artificial intelligence—and which become dependent on others.

Join The Control Layer for weekly perspectives on AI, cybersecurity, and building technology that serves human purpose.

The week that made compute geopolitical

Three announcements in seven days revealed just how explicitly governments now treat AI compute as strategic infrastructure.

On 19 January, the US and Taiwan formalised a $250 billion semiconductor and AI investment pact designed to accelerate the movement of frontier chip production to American soil [1]. The agreement commits Taiwanese semiconductor firms—including TSMC, the world’s most advanced chipmaker—to expand US manufacturing footprint. This is supply-chain decoupling dressed in commercial language.

On 24 January, Singapore’s Ministry of Trade and Industry announced S$1 billion (approximately $779 million USD) in public investment for AI research through 2030, with explicit emphasis on “responsible AI development, compute infrastructure, and industrial deployment” [2]. For a city-state of 5.5 million people, this represents extraordinary per-capita commitment to compute sovereignty.

On 30 January, Reuters reported that Nvidia’s planned $100 billion investment in OpenAI—critical to OpenAI’s training and inference scaling ambitions—had stalled amid internal resistance over deal structure and capital efficiency [3]. Even the firms betting billions on AI are questioning whether the economics hold.

The message is unmistakable: compute is no longer a commodity to be purchased on the open market. It is strategic infrastructure to be secured, subsidised, and protected.

Why compute determines AI power

The AI systems making headlines—GPT-5.2, Claude Opus 4.5, DeepSeek, Qwen—are fundamentally expressions of compute. Training frontier models requires tens of thousands of specialised processors running for months. Inference at scale requires data centres consuming as much electricity as small cities.

Control compute, and you control who can build frontier AI systems. Control the supply chain that produces compute, and you control the future of the technology.

This explains why the US-Taiwan pact matters beyond the headline figure. TSMC manufactures over 90% of the world’s most advanced semiconductors. Its facilities in Taiwan sit 180 kilometres from mainland China. The $250 billion commitment is insurance against geopolitical disruption—ensuring that even if Taiwan’s facilities became inaccessible, the US would retain domestic manufacturing capability for the chips that power AI [1].

Singapore’s investment follows different logic but reaches the same conclusion. As a trade-dependent nation with no natural resources, Singapore has historically survived by positioning itself at critical nodes of global commerce. The S$1 billion AI commitment positions the city-state as a neutral compute hub for Asian enterprises wary of both US and Chinese infrastructure dependencies [2].

The UK’s compute position

Where does this leave Britain?

The UK’s chief sovereign compute asset is Isambard-AI, hosted at the National Compositing Centre in Bristol. At 21 exaFLOPs of AI compute capacity, 217 petaFLOPs of conventional processing, and 25 petabytes of storage, it represents genuine frontier capability [4]. When it came online, it was the most powerful AI supercomputer in Europe.

But capability and strategy are different things. The UK has not matched the US or Singapore with comparable public investment commitments. The anticipated AI Bill did not materialise in 2025 and remains unlikely in 2026 [5]. The government continues its sector-led regulatory approach while other nations are making explicit industrial policy bets.

This creates asymmetric risk. UK organisations building AI capabilities on American cloud infrastructure—AWS, Azure, Google Cloud—remain subject to the US CLOUD Act, which permits US authorities to compel disclosure of data stored by US-headquartered providers regardless of where that data physically resides [6]. For defence contractors, regulated financial services, and critical national infrastructure operators, this is not a theoretical concern.

The choice is not between British isolationism and global integration. It is between strategic autonomy and structural dependency. Other nations are choosing autonomy. The UK’s position remains ambiguous.

Analysis: What strategic compute investment looks like

The US-Taiwan pact and Singapore commitment share three characteristics worth noting.

First, they are explicit industrial policy. Neither announcement pretended to be purely commercial. Both governments stated openly that AI compute infrastructure is a national strategic priority requiring public investment, regulatory alignment, and long-term commitment. This clarity enables coordinated action across government, industry, and research institutions.

Second, they address the full stack. The US-Taiwan pact covers chip manufacturing, not just chip purchasing. Singapore’s commitment spans research, infrastructure, and deployment. Both recognise that compute sovereignty requires domestic capability across the value chain—not just the ability to buy finished products from foreign suppliers.

Third, they create optionality. The US is not abandoning Taiwanese manufacturing; it is creating alternatives. Singapore is not rejecting US or Chinese cloud providers; it is building infrastructure that reduces dependency on either. Strategic compute investment is about expanding choices, not limiting them.

The UK’s Isambard-AI represents genuine capability. But a single supercomputer, however powerful, is not a compute strategy. A strategy would address: domestic semiconductor design and manufacturing capability; sovereign cloud infrastructure for regulated sectors; research funding tied to compute access; and clear policy on foreign data access risks.

In my view, the window for the UK to articulate such a strategy is narrowing. The US, China, Singapore, and the EU are making their positions clear. Ambiguity is itself a strategic choice—and not necessarily a wise one.

The private sector is already responding

While governments debate, enterprises are making their own calculations.

Forrester’s January 2026 survey found that 84% of UK IT leaders express concern about geopolitical data access risks [7]. This is not abstract anxiety. It reflects concrete questions: Can we use US cloud providers for sensitive workloads? What happens if US-China tensions disrupt semiconductor supply? How do we maintain operational continuity if our compute infrastructure becomes a geopolitical liability?

The enterprise response is fragmentation. Organisations are hedging—running some workloads on US hyperscalers for cost and capability, while building parallel infrastructure on European or UK-sovereign alternatives for sensitive operations. This is expensive, complex, and inefficient. But it is rational given policy uncertainty.

For mid-market organisations—typically £50M to £500M in revenue—this hedging is particularly burdensome. Large enterprises can afford multi-cloud strategies with dedicated sovereignty compliance teams. Mid-market firms face the same regulatory and geopolitical risks with a fraction of the resources to manage them.

Risks and constraints

The sovereignty argument carries important limitations.

First, complete compute independence is neither achievable nor desirable. No nation—not even the US or China—can produce every component of the AI stack domestically. The goal is strategic autonomy, not economic self-sufficiency. This means reducing critical dependencies and creating alternatives, not eliminating all foreign supply chains.

Second, sovereignty comes with costs. Domestic infrastructure typically costs more than global alternatives optimised for scale. Organisations choosing sovereign compute may sacrifice capability or efficiency. The question is whether the risk reduction justifies the premium—and that calculation differs by sector, data sensitivity, and threat model.

Third, the geopolitical situation remains fluid. US-UK intelligence cooperation is deep. The CLOUD Act’s practical impact on UK organisations depends heavily on enforcement patterns that could change with administrations. Building strategy around today’s geopolitical alignment may prove shortsighted if that alignment shifts.

Fourth, the UK has genuine strengths. British AI research remains world-class. DeepMind (Alphabet-owned but London-headquartered) and the Alan Turing Institute represent significant intellectual capital. The question is whether these strengths translate into industrial capability or remain vulnerable to acquisition and relocation.

What to do next

For boards and executives: Audit your compute dependencies. Map which workloads run on which infrastructure, with what data residency implications. Understand your exposure to the CLOUD Act and equivalent foreign data access provisions. This is not a technology question—it is a board-level risk question.

For technical leaders: Evaluate sovereign alternatives for sensitive workloads. UK-sovereign cloud providers exist, though with capability trade-offs. Hybrid architectures—US hyperscalers for non-sensitive compute, sovereign infrastructure for regulated data—may represent the practical middle ground while policy clarity emerges.

For mid-market organisations: Engage with industry associations and government consultations on AI and compute strategy. The UK’s approach remains in flux. Organisations that articulate their needs clearly have an opportunity to shape policy. Waiting for clarity may mean accepting terms set by others.

For policy watchers: Monitor DSIT announcements on compute strategy and potential Isambard-AI Phase 2 funding. The government’s response to the US-Taiwan pact and Singapore commitment will signal whether the UK views compute as strategic infrastructure or simply another market to regulate.

Disclaimer: This article represents analysis based on publicly available information as of January 2026. It does not constitute legal, financial, or professional advice.

If your organisation needs support navigating sovereign AI deployment and compliance frameworks, Arkava helps mid-market enterprises build AI capabilities on UK-controlled infrastructure.

References

[1] US Department of State & Taiwan Economic and Trade Office. “US–Taiwan Semiconductor and AI Investment Pact.” 19 January 2026. $250 billion direct investment for compute manufacturing.

[2] Singapore Ministry of Trade & Industry. “National AI Research Initiative (2026–2030).” 24 January 2026. S$1 billion commitment (≈$779m USD).

[3] Wall Street Journal (via Reuters). “Nvidia’s Plan to Invest Up to $100 Billion in OpenAI Has Stalled.” 31 January 2026. https://www.reuters.com/business/nvidias-plan-invest-up-100-billion-openai-has-stalled-wsj-reports-2026-01-31/

[4] The Control Layer (Arkava). “Isambard-AI: The UK’s Leap to the Cutting-Edge.” 23 July 2025.

The Control Layer

Where artificial intelligence, cybersecurity, and enterprise leadership intersect.

By Amer Altaf

— 21 exaFLOPs AI compute; 217 petaFLOPs conventional; 25 petabytes storage.

[5] UK DSIT & Osborne Clarke. “AI Regulation: UK Regulatory Outlook.” January 2026. https://www.osborneclarke.com/insights/regulatory-outlook-january-2026-artificial-intelligence

[6] US CLOUD Act (Clarifying Lawful Overseas Use of Data Act). 2018. Permits US authorities to compel disclosure from US-headquartered providers regardless of data location.

[7] Forrester Research. “Enterprise AI Adoption: Balancing Innovation and ROI in 2026.” 26 January 2026. https://bizzdesign.com/blog/enterprise-ai-adoption-balancing-innovation-and-roi-2026

Why it matters: The EU just handed UK organisations an 18-month window before high-risk AI rules take effect. This isn't compliance relief — it's a strategic inflection point. With the UK government silent on dedicated AI legislation, British businesses must decide alone whether to align with Brussels or chart a distinctively sovereign path.

Join The Control Layer for weekly perspectives on AI, cybersecurity, and building technology that serves human purpose.

The delay nobody expected

Last month, the European Commission proposed something remarkable: pushing back the EU AI Act’s most demanding requirements by more than a year. High-risk AI rules originally due in August 2026 now target December 2027. Systems covered under Annex I face enforcement from August 2028, not 2027.[6]

For organisations that spent 2024 and 2025 scrambling to understand compliance obligations, this feels like an unexpected reprieve. The Brussels machinery, usually relentless in its regulatory march, has effectively admitted it needs more time to develop the technical standards and implementation guidance that make the Act workable.

But reprieve implies you were on a fixed path. UK organisations face a more interesting question: which path are we actually on?

Britain’s curious silence

Here is where the story gets properly British — meaning awkward and unresolved.

The UK government’s response to its own copyright and AI consultation is now due 18 March 2026.[5] That consultation revealed a striking disconnect: only 3% of respondents supported the government’s preferred approach of allowing AI training on copyrighted works with an opt-out mechanism. The majority favoured mandatory licensing. The government has delayed its response twice already.

More telling is what hasn’t happened. Despite months of speculation, no dedicated AI Bill appeared in 2025. AI minister Kanishka Narayan has stated that existing frameworks — data protection, competition law, equality legislation, online safety rules — already apply to AI systems.[5]
The message is clear: we’re not building new regulatory architecture. We’re retrofitting what we have.

A dedicated UK AI Bill is now expected in the second half of 2026 at the earliest. Industry observers consider even that timeline uncertain.[5]

This creates a peculiar situation. The EU is delaying enforcement but accelerating standard-setting — voluntary codes of practice, technical standards committees, implementation guidance flowing from Brussels throughout 2026. The UK is doing neither. No enforcement timeline to slip. No standards to develop. Just silence.

The vacuum is the strategy

Let me be direct about what this means for UK organisations.

You are not waiting for regulatory clarity. You are operating in a deliberate vacuum. The government has made a calculated bet that sectoral regulators — the FCA, Ofcom, the ICO, the CMA — can handle AI governance through existing powers without primary legislation. Whether that bet pays off depends entirely on how those regulators interpret their mandates.

For organisations serving European markets, the calculus is straightforward. EU AI Act compliance remains necessary regardless of UK domestic policy. The delay simply extends your preparation runway. Build frameworks now, refine through 2026, implement ahead of December 2027 enforcement.

But for organisations primarily serving UK markets — defence contractors, public sector suppliers, domestic financial services, healthcare providers — the question becomes genuinely strategic. Do you:

Option A: Build EU-aligned frameworks anyway. Hedge against potential future UK alignment. Maintain optionality for European market expansion. Accept higher compliance costs now for reduced risk later.

Option B: Build fit-for-purpose UK frameworks. Optimise for current UK regulatory expectations. Reduce compliance overhead. Accept the risk that future UK legislation could require significant rework.

Option C: Wait and see. Defer material investment until the UK position clarifies. Minimise near-term costs. Accept concentrated implementation risk in late 2026 or 2027.

None of these options is obviously correct. That’s rather the point.

What the EU delay actually signals

The Commission’s Digital Omnibus proposal isn’t simply about buying time. Read carefully, it reveals where European regulators expect the hard problems to emerge.

The delay to December 2027 for high-risk rules under Article 6(2) and Annex III acknowledges that technical standards aren’t ready. The European standardisation bodies — CEN, CENELEC, ETSI — need another year to produce guidance that makes compliance assessable.[6]

More interesting is what isn’t delayed. AI-generated content labelling obligations under Article 50 remain targeted for February 2027 for systems predating August 2026.[6] The transparency requirements that matter most for public trust stay on track.

The proposal also abolishes the general AI literacy obligation under Article 4, keeping only specific training requirements for deployers of high-risk systems.[6] Brussels is prioritising depth over breadth — focusing enforcement resources on genuinely high-risk applications rather than attempting universal uplift.

And there’s a notable expansion: processing of special category data for bias detection and correction now applies to all AI systems, not only high-risk ones, subject to strict safeguards.[6] The EU is acknowledging that responsible AI development requires access to sensitive data categories, provided appropriate protections exist.

These details matter because they reveal regulatory intent. The EU isn’t backing away from AI governance. It’s refining the approach based on implementation reality.

The UK’s accidental experiment

Here’s a thought that keeps me up at night, in the way only regulatory philosophy can.

The UK is running an uncontrolled experiment in AI governance. By declining to legislate while the EU extends its timeline, we’re creating a natural comparison. Does sectoral regulation through existing frameworks produce better outcomes than comprehensive primary legislation? We’ll find out.

The optimistic reading: UK regulators, freed from prescriptive legislation, can respond flexibly to specific sectoral risks. The ICO addresses AI privacy concerns through GDPR enforcement. The FCA handles financial services AI through existing conduct rules. The CMA examines AI market concentration through competition powers. Each regulator applies domain expertise without the rigidity of a unified AI Act.

The pessimistic reading: without clear primary legislation, UK organisations face fragmented, unpredictable enforcement. Different regulators apply different standards. Gaps emerge between sectoral boundaries. International partners question UK AI governance credibility. Equivalence discussions with the EU become complicated.

Both readings contain truth. The outcome depends on execution — by regulators and by organisations.

What boards should actually do

Let me translate this into actionable guidance, because strategic ambiguity only goes so far.

For boards with EU market exposure:

The December 2027 deadline should anchor your planning. Begin compliance assessment now. Use 2026 to develop classification methodologies, risk management frameworks, and documentation practices. The delay is runway, not relief.

Watch the voluntary code of practice on AI-generated content marking. The second draft arrives March 2026, final code June 2026.[6] Early adoption signals governance maturity to European partners and customers.

For boards serving primarily UK markets:

Map your AI systems against existing UK regulatory frameworks. Identify which sectoral regulators have jurisdiction. Engage proactively — the regulators developing their approaches now will welcome industry input.

Build governance frameworks that work regardless of future legislation. Focus on demonstrable accountability, documented decision-making, and measurable outcomes. Good governance doesn’t require specific regulation.

For all boards:

Don’t conflate regulatory delay with reduced risk. The operational, reputational, and ethical risks from AI systems exist independent of enforcement timelines. Governance frameworks should address actual risk, not compliance calendars.

Use the 18-month window to build institutional capability. Train your teams. Develop assessment methodologies. Create documentation practices. When enforcement eventually arrives — whether from Brussels, Westminster, or both — organisations with mature governance will adapt faster than those starting from scratch.

Analysis: the sovereignty question nobody’s asking

Here’s my honest assessment of where this leads.

The UK’s silence on AI legislation isn’t strategic ambiguity. It’s strategic avoidance.

The government has decided that AI governance is too politically contested, technically complex, and economically uncertain to address through primary legislation in this Parliament. The “pro-innovation” framing provides cover for inaction.

This might work. Sectoral regulation might prove sufficiently flexible and robust. UK AI governance might emerge as a genuine alternative to the EU’s comprehensive approach — lighter touch, more adaptive, better suited to fast-moving technology.

But it might not. The EU is building institutional infrastructure for AI governance: technical standards, certification bodies, enforcement mechanisms, legal precedent. The UK is building none of this. If EU AI Act equivalence becomes a requirement for data flows or market access — not impossible given the trajectory of EU digital policy — the UK will need to construct regulatory architecture rapidly from a standing start.

The 18 months the EU has granted itself is also 18 months for UK policy to clarify.
I would bet significant money that it won’t. British organisations should plan accordingly.

Disclaimer: This article represents analysis based on publicly available information as of January 2026. It does not constitute legal, financial, or professional advice. Organisations should consult qualified advisors for specific compliance guidance.

If your organisation needs support navigating AI governance frameworks — whether aligned with EU requirements or tailored for UK markets — Arkava helps mid-market enterprises turn regulatory uncertainty into strategic advantage.

References

[1] Osborne Clarke. “Artificial Intelligence | UK Regulatory Outlook January 2026.” osborneclarke.com, 12 January 2026. https://www.osborneclarke.com/insights/regulatory-outlook-january-2026-artificial-intelligence

[2] European Commission. “Digital Omnibus Proposal.” November 2025.

[3] UK Government. “AI Regulation: A Pro-Innovation Approach — Policy Proposals.” gov.uk. https://www.gov.uk/government/consultations/ai-regulation-a-pro-innovation-approach-policy-proposals

Why it matters: After years of experimental pilots and inflated promises, agentic AI is producing measurable results in production environments – but only where organisations embed it into tightly scoped workflows with proper governance. The difference between 3:1 ROI and wasted investment now hinges on architecture decisions being made this quarter.

From chatbot curiosity to workflow orchestration

Something shifted in the past twelve months. The AI conversation moved from “what can this chatbot do?” to “how do we orchestrate entire processes end-to-end?” The evidence is now impossible to ignore.

In oil and gas exploration, a platform called NeoSpatial uses large language models to translate natural-language objectives into sequences of geospatial operations, automating workflows that previously required specialist GIS expertise for every scenario.[5] In cardiac catheterisation units, an AI voice assistant named Sofiya completed 806 pre-procedural patient calls over ninety days, handling standardised instructions while nurses focused on complex clinical work.[7] At the Natural History Museum and Royal Botanic Garden Edinburgh, computer vision systems now flag imaging errors in herbarium digitisation at rates of 600 to 1,000 sheets daily, catching mistakes before they trigger costly re-digitisation.[3]

These are not laboratory demonstrations. They are production systems with quantified throughput, measured error rates, and documented governance frameworks.

The architecture that actually works

What distinguishes these successful deployments from the graveyard of abandoned AI pilots? Three architectural patterns appear consistently.

First, narrow initial scope. Sofiya handles pre-procedure calls – not general patient communication. NeoSpatial targets well placement planning, infrastructure routing, and basin evaluation – not open-ended exploration strategy. The NHM-RBGE system detects missing barcodes, cropping errors, and focus problems – not subjective curatorial judgements.[3][5][7]

Second, integrated measurement. Every successful case instruments outcomes from day one. The cardiac catheterisation study tracked call completion rates and patient satisfaction across the full ninety-day period.[7] Enterprise co-pilot programmes measuring task completion times, throughput, and error rates report ROI ranging from 3:1 to 10:1 – while those deploying without baseline metrics struggle to demonstrate value.[4]

Third, explicit governance guardrails. NeoSpatial operates within policy-driven constraints that ensure outputs respect exploration guidelines and regulatory requirements.[5] Radiology workflows embedding agentic AI define complexity thresholds for automation and retain human oversight for high-impact decisions.[6] The lesson is consistent: governance is not a constraint on value – it is a prerequisite for it.

The 3.5x multiplier mid-market firms cannot ignore

Perhaps the most striking finding from recent enterprise deployments concerns how AI co-pilots are positioned within organisations. Firms treating co-pilots as orchestration layers connecting multiple systems – CRM, ticketing, development tools, analytics platforms – achieve approximately 3.5 times the productivity improvement of those rolling out isolated assistants.[4]

The mathematics are instructive. A standalone AI chatbot answering questions about company policy might save an employee five minutes per query. An orchestrated co-pilot that pulls customer history from one system, checks inventory in another, drafts a response, and logs the interaction produces compound efficiency gains across an entire workflow.

For mid-market organisations, this insight reframes the investment question. The cost of AI tooling matters less than the cost of not integrating it properly. A £50,000 co-pilot deployment that remains siloed may deliver less value than a £100,000 implementation wired into existing operational systems with clear measurement from the start.

The control layer insight: Organisations achieving the highest returns treat AI not as a tool to be added, but as an orchestration capability to be architected into existing workflows.

Healthcare as the proving ground

Healthcare offers the most instructive test case because the stakes are highest and the governance requirements most stringent.

The Sofiya deployment demonstrates what responsible automation looks like in practice. The AI voice assistant delivered standardised pre-procedure instructions, collected allergy and medication information, and answered common questions – but all transcripts were recorded and reviewed, with clear escalation paths where the assistant could not safely resolve queries.[7] Nurses remained firmly in the loop. The system handled routine communication; humans handled exceptions and clinical judgement.

Meanwhile, radiology departments are pushing further into multi-agent orchestration. A January 2026 review describes systems where multiple AI agents coordinate protocol selection before imaging, initial image analysis, tool invocation, and preliminary report generation.[6] These configurations outperform single-agent approaches on complex, multi-step processes and reduce manual interventions – but implementation requires defining complexity thresholds, ensuring economic sustainability, and building appropriate governance frameworks.

The pattern generalises beyond healthcare. Automation works when it addresses clearly bounded tasks with measurable outcomes and human oversight for edge cases. It fails when organisations attempt to automate judgement before they have automated routine.

Analysis

The shift from experimental AI to production agentic systems marks a genuine inflection point – but not for the reasons typically cited in technology commentary.

The underlying models have not suddenly become dramatically more capable. What changed is organisational maturity. The early adopters who invested in data infrastructure, workflow documentation, and governance frameworks three years ago now have the foundations to deploy orchestrated AI systems that actually deliver. Organisations that skipped those foundations are discovering that bolting AI onto broken processes produces broken AI.

This creates a widening gap. Firms with instrumented, governed workflows can layer agentic automation and measure results within quarters. Those without are stuck in pilot purgatory, unable to demonstrate value because they lack the baseline data to prove improvement.

For boards evaluating AI investments, the implication is uncomfortable but clear: the biggest barrier to ROI is not technology cost or model capability. It is the absence of the operational discipline required to deploy technology responsibly.

Risks and constraints

Three significant constraints warrant attention.

The measurement challenge cuts both ways. Organisations that instrument their workflows can demonstrate value, but they also expose themselves to evidence that AI investments are not delivering. Some leaders may prefer the ambiguity of unmeasured pilots to the accountability of documented results. This is a cultural problem that technology cannot solve.

Governance frameworks remain immature. While NeoSpatial and radiology deployments describe policy guardrails and complexity thresholds, industry-wide standards for agentic AI governance are still emerging. Organisations deploying today must develop bespoke governance approaches, with the attendant risk that their frameworks may not align with regulations that arrive later.

Talent constraints persist. Successful implementations require people who understand both the business domain and the technical architecture of agentic systems. The NeoSpatial case explicitly notes reduced dependency on scarce GIS specialists – but the platform itself required substantial expertise to build.[5] The skills gap has shifted, not disappeared.

What to do next

For boards and executives: Demand baseline metrics before approving AI investments. If your organisation cannot measure current workflow performance, it cannot measure AI-driven improvement. The first investment should be in instrumentation, not automation.

For technical leaders: Evaluate AI opportunities through an orchestration lens. Ask not “what can this tool do?” but “how does this integrate with our existing systems and data flows?” Prioritise implementations that connect multiple systems over those that operate in isolation.

For mid-market organisations: Start with your most clearly documented, highest-volume workflow. Scope tightly, measure rigorously, and build governance from day one. The 3.5x multiplier for integrated co-pilots suggests the competitive advantage lies not in being first to deploy AI, but in being first to deploy it properly.

Disclaimer: This article represents analysis based on publicly available information as of January 2026. It does not constitute legal, financial, or professional advice.

If your organisation needs support implementing AI governance frameworks that deliver measurable outcomes, Arkava helps mid-market enterprises turn AI investment into accountable results. Contact: engage@arkava.ai

References

[1] UK Government Digital Service. “UK government roadmap outlines next steps for digital transformation.” Global Government Forum, 21 January 2026. https://www.globalgovernmentforum.com/uk-government-roadmap-unveils-next-steps-in-modernising-digital-government/

[3] Poon et al. “AI-Assisted Quality Checks for Herbarium Sheet Digitisation.” Biodiversity Information Science and Standards, 13 January 2026. https://biss.pensoft.net/article/184058/

[4] Troylendman. “AI Co-Pilot Adoption 2025: Transforming Work Through Case Studies.” July 2025. https://troylendman.com/ai-co-pilot-adoption-2025-transforming-work-through-case-studies/

[5] IPTC. “NeoSpatial: Agentic AI Driven Solution, Revolutionizing Oil and Gas Exploration Geospatial Workflows.” January 2026. https://onepetro.org/IPTCONF/proceedings/26IPTC/26IPTC/D011S009R006/794830

[6] Radiology: Artificial Intelligence. “Agentic AI in Radiology: Evolution from Large Language Models to Future Clinical Integration.” January 2026. http://pubs.rsna.org/doi/10.1148/ryai.250651

[7] European Heart Journal – Digital Health. “Utilizing an AI-assisted virtual agent for pre-procedural patient calling in the cardiac catheterization laboratory.” 31 December 2025. https://academic.oup.com/ehjdh/article/doi/10.1093/ehjdh/ztaf143.007/8422867

Why it matters: After years of AI summits producing safety frameworks and principles, India is hosting one explicitly focused on "impact" – measurable outcomes that justify investment. For organisations struggling to demonstrate AI ROI, this signals a broader shift in how governments and industry will evaluate AI success: not by capability, but by value delivered.

The expensive conversation we keep having

AI summits have become a reliable feature of the international calendar. The UK hosted one in 2023. Korea followed. France had its turn. Each produced communiqués, frameworks, and commitments to safety principles that governments could reference in press releases.

What they have not produced, with any consistency, is a framework for measuring whether AI investments actually deliver value.

India’s upcoming “Impact Summit” represents an explicit break from this pattern. The name itself is a statement: not safety, not innovation, not ethics. Impact. Measurable outcomes that matter to people and organisations.

For UK boards watching AI budgets grow while ROI remains elusive, this shift in global conversation matters more than any safety framework.

The value gap is real

Consider the operating reality for most organisations. AI spending is increasing. Vendor promises are impressive. Pilot projects proliferate. And finance directors keep asking the same uncomfortable question: what are we actually getting for this?

Industry estimates suggest >75% of AI investments fail to deliver measurable value [Analysis based on industry benchmarks]. That is not a technology problem. It is a value problem. [2]

India’s High Commissioner to the UK, Vikram Doraiswami, framed this directly at TechUK: “It should not only be about spending money on AI. It should be about what you are trying to do and who will it benefit.”

This is a different starting point than most AI governance conversations. Not “how do we ensure AI is safe?” but “how do we ensure AI is worth it?”

“It should not only be about spending money on AI. It should be about what you’re trying to do and who will it benefit.” – H.E. Vikram Doraiswami [1]

The space programme lesson

Doraiswami offered an instructive parallel from India’s history: the space programme.

In the 1960s, India investing in space technology was widely mocked. Why does a country with such poverty need satellites? There were images of rocket components transported on bullock carts – an easy visual for those who thought the investment absurd.

India’s answer was practical: what can new technologies do for us in developmental terms?

The space programme was designed for immediate returns. Remote sensing provided information for farming and land-use patterns. As capabilities improved, meteorology inputs reduced the impact of cyclones and tidal surges. Losses from climatic catastrophes are now a fraction of what they were 30 or 40 years ago – directly attributable to space programme investments.

The cost efficiency became legendary. India’s 2023 moon landing on the lunar south pole reportedly cost $75 million end-to-end. Alfonso Cuarón’s film Gravity cost $100 million. A Hollywood movie, literally more expensive than a moon landing.

Analysis: This is not about celebrating frugality for its own sake. It is about demonstrating that technology investment can be evaluated on outcome-to-cost ratios rather than capability alone. The moon landing was not impressive because it was cheap. It was impressive because it achieved a specific objective at minimal resource expenditure.

What “impact” means in practice

The Impact Summit is organised around seven verticals, according to Doraiswami, each bringing together government, academic, and business partners. The structure deliberately avoids the traditional summit format of segregated discussions – ministers in one room, academics in another, businesses in a third.

Instead: regulators, innovators, and deployers of AI systems in the same conversations, focused on specific use cases.

This is not conceptually revolutionary. But it is structurally different from summits designed to produce principles rather than implementations.

The verticals include “Safe and Trusted AI” – covering standards for safety testing, auditing, governance, and transparency frameworks. But notably, this is one pillar among seven, not the entire agenda. Safety is necessary. It is not sufficient.

The multilingual proof point

If you want to understand what “impact-first” AI looks like in practice, consider India’s approach to language.

India has 22 officially recognised languages. English speakers number around 140 million – only 10% of the population [1]. An AI strategy that works only in English is, by definition, an AI strategy that excludes 90% of citizens.

The response has been to train AI models across languages. According to Doraiswami, 22 Indian languages now have functional AI solutions. The Bhashini platform provides translation of government and legislative documents. Court judgements – previously accessible only in English or Hindi – can now be read across the country in citizens’ preferred languages,

One concrete example: Marathi-language AI farming software reportedly serves 16 million users in a state of approximately 110 million people. [3] Specialised inputs about farming, accessible in the language farmers actually speak.

This is not AI for AI’s sake. It is AI evaluated on whether it reaches people who need it, in forms they can use.

Analysis: For UK organisations, the lesson is not about multilingual deployment (though that matters for diverse populations). It is about the evaluation framework. India is asking: does this AI investment improve outcomes for intended users? That question should precede any conversation about safety, ethics, or capability.

The interoperability imperative

There is a second dimension to the impact agenda: interoperability.

Doraiswami drew an analogy to email protocols. Forty years ago, SMTP created an open, common railroad for electronic communication. Different providers, different interfaces, universal connectivity. Everyone could email everyone else.

Then came WhatsApp, Signal, iMessage. Walled gardens where interoperability declined. Convenient, but fragmented.

The question for AI: are we building SMTP, or are we building walled gardens?

India’s preference, expressed through its digital public infrastructure model, is clear: shared rails where everyone can build applications. Government owns the infrastructure. Private sector innovates on top. Competition happens at the application layer, not the platform layer.

If this sounds abstract, consider the UK mid-market organisation evaluating AI governance tools. The current choice is typically between expensive enterprise platforms (often US-controlled, rarely interoperable) and DIY approaches (cheap but fragmented). Neither resembles infrastructure designed for shared benefit.

The Impact Summit’s agenda includes discussions about “globally subsidised GPU access” – treating compute as shared strategic infrastructure rather than competitive advantage. Whether this is politically feasible is questionable. That it is being discussed at a major summit is notable.

What the UK-India partnership could mean

The joint India-UK AI centre announced during Prime Minister Starmer’s October 2024 visit could operationalise this agenda.

The proposed model: testing sandboxes where governance frameworks can be prototyped and validated before deployment. Not regulation designed in abstraction, but infrastructure tested against real use cases.

For UK mid-market organisations, this could matter significantly. Participating in sandbox testing provides access to governance frameworks developed collaboratively – without the cost of building proprietary compliance infrastructure.

The value proposition is straightforward: governance as shared infrastructure, reducing the compliance burden for organisations that cannot afford enterprise-scale investment.

Analysis: The question is whether UK government will lean into this opportunity or treat the partnership as diplomatic theatre. The difference will be visible in funding, participation structures, and whether sandbox outputs become accessible to organisations beyond those directly involved.

Risks and constraints

The impact agenda is not without significant challenges.

Verification matters. Many statistics cited like the 16 million Marathi AI users come from government sources with incentive to present favourable narratives. Independent verification is essential before treating these as established benchmarks.

Impact can be defined self-servingly. Any government can claim its programmes have impact. The question is whether impact is measured against objectives that matter to citizens, or objectives that make programmes look successful. Rigorous, independent evaluation frameworks are harder to build than principles documents.

Cultural and institutional differences. What works in India’s governance context – strong central coordination, developmental imperatives, different privacy expectations – may not transfer directly to UK contexts. Learning from the model requires adapting it, not copying it.

The Global South framing cuts both ways. India positions itself as speaking for developing nations. But India is also a nuclear power, space power, and aspiring superpower. Its interests do not automatically align with smaller developing countries, and “Global South” solidarity should be evaluated critically.

Energy constraints remain real. Doraiswami acknowledged that renewable energy has limits, particularly around storage. India is pursuing green hydrogen and nuclear expansion to support AI compute growth. Without solving the energy question, AI ambitions – however well-intentioned – face hard physical constraints.

What to do next

For boards and executives:

• Before your next AI investment decision, require a clear answer to: “What specific outcome are we trying to achieve, and how will we measure whether we achieved it?”

• Benchmark AI investments not against peer spending, but against value delivered per pound invested. India’s space programme model – capability achieved relative to resources expended – is instructive.

For technical leaders:

• Evaluate your AI governance approach against interoperability criteria. Can your frameworks work with others, or do they create lock-in?

• Watch the Impact Summit outcomes for sandbox opportunities and governance infrastructure that could reduce your build-versus-buy burden.

For mid-market organisations:

• Track the UK-India AI centre developments. Early participation in testing sandboxes could provide governance frameworks without enterprise-scale investment.

• Challenge vendors to demonstrate impact metrics, not capability metrics. The question is not what the AI can do. The question is what it will deliver for your specific context.

Disclaimer: This article represents analysis based on publicly available statements from a TechUK event in January 2025. Statistics cited are attributed to the speaker and require independent verification. This does not constitute legal, financial, or professional advice.

If your organisation needs support building AI governance frameworks that deliver measurable outcomes – not just compliance checkboxes – Arkava helps mid-market enterprises turn AI investment into business value.

Contact: engage@arkava.ai

References

[1] H.E. Vikram Doraiswami. Interview at TechUK, January 2025. Transcript summary provided.

[2] IBM Study: 75% of AI Projects Fail
[3] Primary Census Abstract (PCA) 2011 PDF, hosted on the official Maharashtra Directorate of Economics and Statistics site