BolgiaTen Blog

Why AI transformation now belongs on the CEO agenda – and why the decisive advantage will be human, not merely technical

The argument

Over the past few years, the AI conversation in business has moved through three distinct phases. The first was fascination: generative AI as an extraordinary instrument for writing, searching, summarising and coding. The second was experimentation: pilots, sandboxes, copilots, innovation days and executive theatre. The third phase, the one now arriving, is more serious. AI is becoming a test of organisational speed. Not speed as haste, not speed as uncontrolled adoption, and certainly not speed as buying every fashionable tool in the market. I mean the deeper speed of an enterprise: the capacity to sense change, decide intelligently, redesign work, govern risk and learn faster than the environment is changing around it.

This matters particularly for large multinational enterprises carrying accumulated technical debt in tools, infrastructure, data and management practice. Many of these organisations were built for scale, resilience and control, not for continuous recomposition. Their ERP estates, data lakes, cloud migrations, procurement cycles, cyber controls, risk committees and legacy applications were not designed for a world in which intelligence is becoming embedded in every workflow. AI has exposed what was already true: the limiting factor is rarely the model. The limiting factor is the operating model.

This is the point I have tried to make in my recent writing on Generative AI and Agentic AI. The interesting question is no longer whether a model can produce an acceptable answer. The question is whether the organisation can turn that answer into a governed action, at scale, in context, with accountability. Agentic AI intensifies the issue because it shifts the discussion from tools that assist people to systems that initiate, plan, call other systems, execute tasks and learn from feedback. That is not a software upgrade. It is a challenge to the organisation’s metabolism.

What has changed

The empirical evidence now shows two truths moving together. First, adoption has accelerated dramatically. Stanford’s 2025 AI Index reported that 78% of organisations were using AI in 2024, up from 55% the previous year, while generative AI investment continued to expand significantly. McKinsey’s 2025 State of AI survey similarly describes wider use of AI and agentic AI, but also notes that many organisations are still struggling to move from pilots to scaled economic value. The pattern is clear: AI has crossed the adoption threshold, but not yet the transformation threshold.

Second, we are learning that value does not arrive evenly. Brynjolfsson, Li and Raymond’s research on generative AI in customer support found average productivity improvements of about 14%, with the largest gains accruing to less experienced workers. Dell’Acqua and colleagues, in their study with BCG consultants, described the ‘jagged technological frontier’: AI can lift performance significantly for tasks within its frontier and degrade performance for tasks outside it. This is crucial for boards. AI is not a universal accelerator. It is a conditional accelerator. It rewards judgement, task decomposition, good data, domain context and feedback. It punishes blind delegation.

This is why so many pilots disappoint. MIT’s 2025 GenAI Divide report argued that many enterprise initiatives fail because they are brittle, poorly integrated into daily work and unable to learn from context. Deloitte’s 2025 enterprise research similarly points to rising investment alongside elusive returns. IBM’s 2025 CEO research found that rapid investment has often created disconnected technology, while IBM’s 2026 CEO research reported that 83% of CEOs believe AI success depends more on people adoption than on technology itself. The message is no longer subtle: AI transformation fails when it is treated as deployment rather than rewiring.

From digital transformation to intelligent transformation

For thirty years, business transformation was largely about digitising existing processes. We put forms online, moved workloads to cloud, integrated channels, automated back offices and introduced analytics. Much of this was valuable, but it often preserved the inherited shape of the organisation. AI is different because it changes the unit of work. It can read, reason, generate, classify, converse, code and increasingly orchestrate. In agentic form, it can become a new participant in the enterprise operating system.

That creates a dangerous temptation: to insert AI into old processes and call it transformation. A CEO with a heavily indebted technology estate should resist this. If the process is broken, AI will accelerate the brokenness. If the data is fragmented, AI will make the fragmentation visible. If accountability is unclear, AI will amplify ambiguity. If middle management has been trained to protect functional boundaries, AI will not magically create cross-enterprise flow. The organisation will merely become faster at revealing its own incoherence.

The better question is: where are the enterprise constraints that AI now makes negotiable? Which approvals exist because information used to be scarce? Which reports exist because systems could not explain themselves? Which roles exist to reconcile data that should never have been inconsistent? Which customer journeys are slow because the organisation is divided by internal functions rather than external outcomes? Which technical debt has been tolerated because the cost of change was historically too high? AI changes the cost curve of coordination, but only if leadership is willing to challenge the contracts embedded in the organisation.

What it means to build organisational speed

Organisational speed is not the same as moving quickly. Many organisations are already fast in the wrong places. They can launch pilots quickly, buy tools quickly and issue press releases quickly. The more valuable form of speed has five characteristics.

The first is speed of sense-making. Leaders need the ability to detect where AI is changing customer expectations, cost structures, risk profiles and competitive boundaries. This requires external scanning, internal telemetry and board-level fluency. A board that treats AI as a technology topic will be late; a board that treats AI as a strategic discontinuity has a chance.

The second is speed of decision. AI opportunities decay when they are trapped in committees designed for yesterday’s risk. This does not mean weakening governance. It means designing governance that is proportionate, informed and close to the work. Responsible AI, security, data protection and model assurance must be built into the delivery system, not bolted on as a final inspection.

The third is speed of learning. Organisations must move from pilot culture to learning culture. A pilot asks whether a tool works. A learning system asks what changed in the work, what was adopted by people, what risk emerged, what data improved, what should be stopped and what should scale. This is where many enterprises are weakest. They accumulate experiments without compounding knowledge.

The fourth is speed of integration. The next advantage will not come from isolated copilots. It will come from connecting models to workflows, data, controls, APIs, human review, cyber policy, auditability and business outcomes. This is where technical debt becomes strategic debt. Legacy infrastructure is not merely an IT inconvenience; it is a brake on organisational learning.

The fifth is speed of trust. People will not adopt systems they do not understand, cannot challenge or believe are being used against them. Trust is not soft. It is the lubricant of transformation. Without it, employees route around new tools, managers preserve old behaviours and the organisation creates a theatre of adoption while real work continues elsewhere.

Why the deepest transformations are about people

BCG has often framed AI value through a 10-20-70 logic: a smaller proportion lies in algorithms, more in technology and data, and the majority in people, process and change. Whether one accepts the exact numbers or not, the principle is right. The transformation is ultimately human because work is social before it is technical. Decisions are made by people, exceptions are handled by people, customers trust people, risk is owned by people and culture is transmitted by people.

The World Economic Forum’s Future of Jobs Report 2025 expects 39% of workers’ core skills to change by 2030 and estimates that, in a workforce of 100 people, 59 will need training before the end of the decade. That is not an HR footnote. It is a balance sheet issue. Skills are now a strategic asset class. The enterprise that cannot reskill quickly cannot transform quickly. The enterprise that cannot redesign roles cannot capture AI value. The enterprise that treats people as recipients of change rather than authors of change will lose the very intelligence it needs.

This is particularly true for middle management. In many large enterprises, middle managers are the translation layer between strategy and work. They can either become the accelerators of AI transformation or its immune system. If they are excluded, threatened or left untrained, they will slow the transformation in rational self-defence. If they are equipped to redesign work, coach teams, manage risk and interpret AI outputs, they become the most important agents of speed.

The same is true for frontline expertise. AI systems require context. They need to learn from the people who know where processes fail, where customers become frustrated, where data is misleading, where policies contradict reality and where exceptions actually occur. In this sense, AI does not remove the need for human intelligence; it increases the premium on human judgement. The future enterprise is not a machine with people attached. It is a human institution with new cognitive infrastructure.

The CEO agenda

For the CEO of a large multinational enterprise, the practical implications are stark. First, do not allow AI to become another layer of technical debt. Every AI investment should be tested against architecture, data lineage, cyber posture, model governance and integration into real work. Second, move from use-case enthusiasm to capability building. The question is not how many pilots are running, but whether the organisation is building reusable data products, model assurance, workflow orchestration, talent pathways and decision rights. Third, make adoption a leadership discipline. Usage statistics are not enough; measure changes in cycle time, quality, customer outcomes, employee confidence and risk controls.

Fourth, create a strategic map of what must be rewired. Some processes should be automated, some augmented, some eliminated and some protected because human judgement is the source of value. Fifth, put people at the centre without romanticising the status quo. People-led transformation does not mean avoiding difficult choices. It means making those choices with clarity, fairness, participation and investment in capability.

My own view is that AI transformation is entering its second act. The first act was about possibility. The second is about organisational character. The winners will not be the firms with the most pilots, the largest tool catalogue or the loudest AI narrative. They will be the firms that build speed without losing judgement, automate without abandoning accountability, and use AI to enlarge human agency rather than merely reduce human cost.

That is why this is a CEO issue. Technical debt, process debt and skills debt have converged. AI has made the hidden friction of the enterprise visible. The question for the board is not whether the organisation should adopt AI. That decision has already been made by the market. The question is whether the organisation can rewire itself quickly enough, wisely enough and humanely enough to turn intelligence into advantage.

Selected references

Brynjolfsson, E., Li, D. and Raymond, L. R. (2023/2025), ‘Generative AI at Work’, NBER Working Paper 31161 and Quarterly Journal of Economics.

Dell’Acqua, F. et al. (2023/2025), ‘Navigating the Jagged Technological Frontier’, Harvard Business School / Organization Science.

Deloitte (2025), The State of Generative AI in the Enterprise.

IBM Institute for Business Value (2025), CEO Study: CEOs Double Down on AI While Navigating Enterprise Hurdles.

IBM Institute for Business Value (2026), CEO Study: CEOs are Reshaping C-suite Roles for the AI Era.

McKinsey & Company (2025), The State of AI: Global Survey 2025.

MIT NANDA (2025), The GenAI Divide: State of AI in Business 2025.

Stanford HAI (2025), Artificial Intelligence Index Report 2025.

World Economic Forum (2025), The Future of Jobs Report 2025.

Source URLs consulted: 

https://hai.stanford.edu/ai-index/2025-ai-index-report

https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-state-of-ai

https://www.nber.org/papers/w31161

https://www.hbs.edu/faculty/Pages/item.aspx?num=64700

https://www.deloitte.com/uk/en/issues/generative-ai/state-of-generative-ai-in-enterprise.html

https://newsroom.ibm.com/2025-05-06-ibm-study-ceos-double-down-on-ai-while-navigating-enterprise-hurdles

https://newsroom.ibm.com/2026-05-04-ibm-study-ceos-are-reshaping-c-suite-roles-for-the-AI-era

https://mlq.ai/media/quarterly_decks/v0.1_State_of_AI_in_Business_2025_Report.pdf

https://www.weforum.org/publications/the-future-of-jobs-report-2025

Why boards and CEOs must look beyond application-layer productivity and ask which hidden contracts in the business technology stack are about to be renegotiated by Agentic AI

The wrong conversation is winning

In recent posts I have argued that Generative AI is not simply a tool for producing more words, more code, more images or more slideware. I have also argued that Agentic AI is not merely “automation with a better user interface”. Agentic AI changes where agency sits in the organisation. It moves decision, orchestration and execution into systems that can observe, reason, act, test, learn and recover. That is why I keep returning to governance, accountability and economic impact. The technology is interesting, but the transfer of agency is the point.

Most board conversations, however, are still stuck in the most visible part of the game. They ask whether developers will be 20% or 40% faster. They ask whether a copilot can clear the technology backlog. They ask whether the customer service team can answer more tickets with fewer people. These are legitimate questions, but they are scoreboard questions. They tell us whether today’s team is running harder under today’s rules. They do not tell us whether the rules of the sport are being rewritten.

The deeper AI story is not that software teams are being given better boots. It is that the pitch, the rules, the coaching staff, the refereeing system, the medical science, the broadcast model and the business model of the club are all becoming writable at the same time. Boards that only watch the striker will miss the change in the entire league.

From productivity story to paradigm story

For fifty years, the technology stack has behaved like a professional sport with rigid divisions of labour. The players play, coaches coach, referees referee, grounds teams maintain the surface, broadcasters package the spectacle, and owners negotiate commercial rights. Each role has its own contract, language, incentives and power base. In technology, those contracts are the application interface (API’s), the runtime, the operating system, the compiler, the instruction set and the silicon. We have treated them as fixed because changing them together was prohibitively expensive.

The application developer did not negotiate with the chip. The chip designer did not negotiate with the customer journey. The compiler expert did not sit in the boardroom discussing revenue leakage. The contracts were not laws of physics. They were historical settlements that lasted so long they started to look inevitable.

Large language models (LLM’s) and Agentic systems now challenge that assumption. They can work across the boundaries that humans built around expertise. They can write application code, generate compiler optimisations, synthesize kernel extensions, assist chip design, support formal verification and help redesign silicon layouts. They do not respect the old dressing-room politics because they did not grow up inside them. They ask a question most organisations have been structurally unable to ask: what happens if several layers of the game can be changed together?

That is why this is not primarily a productivity story. It is a paradigm story. Productivity improves the existing formation. Paradigm change alters what a formation is.

The evidence is already on the field

This is not science fiction. At the compiler layer, Meta’s LLM Compiler was trained on 546 billion tokens of LLVM intermediate representation and assembly code. Its published results show 77% of the optimisation potential of an autotuning search and 45% disassembly round-trip accuracy from x86 and ARM assembly back into LLVM IR. In plain English, models are learning the language beneath the language most software teams discuss.

At the operating-system layer, Kgent, presented at the ACM SIGCOMM 2024 eBPF workshop, translates natural language prompts into eBPF programs for the Linux kernel. Related 2025 work on agentic operating systems uses LLM agents to analyse workloads, synthesize eBPF scheduling policies and deploy them through sched_ext, reporting up to 1.79x performance improvement and a 13x cost reduction compared with naive agent approaches. The “grounds team” of the computing world is no longer only repairing turf; parts of the turf-management strategy are becoming agentic.

At the chip-design layer, NVIDIA’s ChipNeMo explores domain-adapted large language models for industrial chip design, including engineering assistant use cases, EDA script generation and bug summarisation. Google DeepMind’s AlphaChip uses reinforcement learning to generate chip floorplans in hours rather than the weeks or months traditionally required, and Google says these layouts have been used in multiple generations of TPUs and in Axion CPU work. At the instruction-set layer, RISC-V research on the XiangShan Nanhu processor extends the instruction set for LLM vector dot-product acceleration, reporting more than four times scalar-method speed on that core operation.

The important point is not that any one of these examples wins the match alone. The point is that they are appearing at multiple layers at once. This is the early sign of what I have called consurgence: new properties rising together because the conditions have changed together.

A sporting metaphor for the boardroom

Imagine a football club that has spent decades improving marginally within known constraints. It hires better strikers, buys improved boots, upgrades nutrition, deploys data analysts and improves ticketing. Each initiative is useful. Some are very profitable. But all assume the same pitch, the same league rules, the same broadcast economics and the same match-day model.

Now imagine a new class of agentic system arrives that can redesign the training plan, simulate match tactics, change the playing surface, alter recovery science, generate new scouting models, negotiate media packages, optimise stadium logistics and propose rule changes to the league. The club that asks only, “Can our striker score 10% more goals?” has misunderstood the opportunity. The better question is, “Which assumptions about the game have become negotiable?”

That is where companies are today. The visible AI market is crowded with applications: copilots, assistants, chatbots, content tools, workflow wrappers and SaaS add-ons. This is the striker market. It is exciting, noisy and overfunded. But the deeper advantage will come from the less glamorous parts of the club: the academy, sports science, data infrastructure, playing surface, referee technology, transfer analytics and league governance. In technology terms, those are compilers, runtimes, verification, operating systems, open silicon tooling, capability security and the governance of agentic action.

Why this matters to CEOs and boards

Boards do not need to become chip designers. CEOs do not need to run compiler teams. But they do need to understand where durable advantage is likely to form. If AI simply improves the application layer, then the winning strategy is adoption speed: choose tools, train staff, measure productivity and manage risk. If AI rewrites the layers beneath the application, the winning strategy is different: identify which hidden contracts shape your economics and decide whether to defend, renegotiate or escape them.

For a telco, the hidden contract may be between network operations, customer experience and vendor equipment roadmaps. For a bank, it may be between risk models, legacy core systems, regulatory reporting and the human sign-off process. For a logistics company, it may be between routing, fleet maintenance, insurance, carbon reporting and customer promises. For a public-sector body, it may be between policy intent, operational data, procurement rules and citizen outcomes. Agentic AI is not powerful because it “does tasks”. It is powerful because it can coordinate across these boundaries and reveal which boundaries were artificial.

This is also why AI governance cannot be reduced to ethics theatre or model-risk paperwork. My position has consistently been that governance must be an operating discipline. When agents can act across systems, governance must define permission, accountability, evidence, reversibility and escalation. The question is not only “Was the model fair?” It is also, “Who authorised the agent to change the play, under what constraints, with what audit trail, and who can stop it when the match changes?”

The capital allocation mistake

The current investment pattern is heavily weighted towards the obvious. Most corporate pilots and much venture capital crowd around application-layer tools because they are easy to demonstrate and easy to sell. The board can see the demo. The CFO can imagine the headcount saving. The press release almost writes itself.

But if the real renegotiation is happening below the application layer, then the smarter capital question changes. Companies should still invest in adoption, but not confuse adoption with advantage. The layers that deserve more attention are the systems that translate new compute into usable business capability: compiler and runtime infrastructure, formal verification, secure agent execution, open silicon design tooling, heterogeneous compute operating models, and hardware-enforced capability boundaries such as CHERI and Arm Morello. These are not fashionable board topics, but neither were cloud control planes, mobile app stores or semiconductor supply chains until they determined who captured the margin.

The UK policy picture illustrates the risk. The AI Opportunities Action Plan and subsequent updates rightly emphasise compute, AI Growth Zones, public-sector adoption, data assets, sovereign capability and skills. UKRI’s 2026 AI strategy commits more than £1.6 billion of targeted AI funding, while the Sovereign AI Unit is backed by up to £500 million. These are serious commitments. Yet the public framing still leans towards bedrock compute and visible adoption. The middle layers, where contracts become moats, receive far less public attention.

That is not a criticism of intent. It is a challenge to worldview. If you believe AI is another wave to be added to the existing stack, the current allocation is coherent. If you believe AI makes the stack itself writable, then the allocation looks incomplete.

The board agenda should change

The practical board agenda should now include five questions.

First: which technology contracts have we treated as fixed because changing them was historically too expensive? These may be vendor contracts, data architectures, operating models, platform dependencies, regulatory workflows or pricing models.

Second: where are we confusing productivity with strategic advantage? A 30% faster process inside a soon-to-be-obsolete model is not transformation. It is a better warm-up before the wrong match.

Third: which agents will be allowed to act, not merely advise? The moment an agent can trigger workflow, move money, reconfigure infrastructure, change a customer promise or alter a policy, the governance model must mature from advisory AI to accountable agency.

Fourth: are we investing only in the striker, or also in the academy, pitch, data room and rulebook? The visible applications matter, but they may become commoditised quickly. The less visible layers may determine lock-in, security, resilience and strategic optionality.

Fifth: what would we do differently if the cost of coordinated change fell by an order of magnitude? This is the question that forces leadership to think beyond process improvement and into new operating models.

A different game is forming

The danger for incumbents is not stupidity. It is success. Incumbents are optimised around the current contracts. Their management systems, budgets, partner ecosystems and board reporting all assume the sport remains recognisable. They will naturally describe change as incremental because their advantage is built inside the old rulebook.

History is not kind to those who manage the visible crisis and miss the structural one. Nokia, Kodak, Xerox and Blockbuster were not short of smart people. They were short of timely courage. They read the current scoreboard well and misread the future league.

The same pattern is visible now. Many leaders are debating whether AI can patch the roof of the current software business model: per-seat pricing, SaaS renewals, developer velocity, customer-service automation. Those discussions matter, but they are not enough. Agentic AI points towards a world where the unit of value may not be the application, the seat or even the workflow. It may be the agentic capability to orchestrate outcomes across systems, with governance strong enough to make that capability trusted.

The CEO takeaway

The old question was: how much faster can AI make our people? The better question is: which assumptions in our business have become negotiable?

The old question was: where can we add GenAI to the existing application estate? The better question is: where are agents about to coordinate across boundaries that our organisation chart still treats as sacred?

The old question was: should we invest in AI tools? The better question is: are we investing in the layers where advantage will compound, or only in the layer where the crowd has gathered?

In sport, the great clubs do not only buy stars. They build systems: academies, analytics, coaching, medical capability, recruitment intelligence, stadium economics and culture. They understand that sustainable advantage is rarely found only in the moment the ball hits the net. It is created in the conditions that make that moment repeatable.

That is the lesson for boards and CEOs. Generative AI has made the visible play more productive. Agentic AI is beginning to change who calls the play, how the play is executed, and whether the rules themselves can be rewritten. The scoreboard will not tell you that early enough. You have to look at the whole game.

My advice is simple: stop watching only the striker. The game is changing under the pitch.

Selected references and further reading

• Meta LLM Compiler: Foundation Models of Compiler Optimization

• NVIDIA Research: ChipNeMo – Domain-Adapted LLMs for Chip Design

• Google DeepMind: How AlphaChip transformed computer chip design

• Kgent: Kernel Extensions Large Language Model Agent, ACM SIGCOMM eBPF 2024

• Towards Agentic OS: An LLM Agent Framework for Linux Schedulers

• RISC-V XiangShan Nanhu LLM acceleration research

• UK Government: AI Opportunities Action Plan – One Year On

• UKRI Artificial Intelligence Research and Innovation Strategic Framework

• Arm Morello Program and CHERI capability technology

How national control of compute, data and models will reshape data centre design, location and corporate strategy

The argument

For many years, the phrase “data sovereignty” was treated as a compliance issue: where is the data stored, who can access it, and which regulator has authority over it? AI sovereignty is different. It is about whether a nation can create, operate, audit, secure and adapt the intelligence that will increasingly sit inside public services, critical infrastructure and the balance sheets of private enterprises. In my view, this is not a fashionable policy phrase. It is the next layer of national capability, sitting alongside energy, telecommunications, defence, logistics and financial resilience.

Best practice is already converging around a wider definition. The OECD argues that national AI compute plans need to address capacity, effectiveness and resilience, including “security, sovereignty, sustainability”. [1] The UK Government has committed to expand “sovereign compute capacity by at least 20x by 2030”. [2] The European Commission’s AI Continent plan is even more explicit: AI Gigafactories are intended to train complex models, with up to five facilities mobilised through InvestAI, while the proposed Cloud and AI Development Act aims to triple EU data centre capacity in the next five to seven years. [3]

This changes the question for governments and companies. Sovereignty is no longer achieved by placing a server in a capital city and declaring victory. It requires control over the stack: data, connectivity, chips, energy, cooling, model governance, cyber resilience, jurisdiction, procurement and the ability to operate under stress.

From centralised cloud to sovereign AI fabric

The early cloud model favoured concentration: a few hyperscale regions, massive efficiency, global platforms and standardised operating models. AI will not abolish that model, but it will stretch it. Training frontier models may still require enormous specialised clusters. Inference – the daily running of AI inside customer service, fraud detection, logistics routing, insurance pricing or network operations – will be persistent, distributed and latency-sensitive.

That distinction matters. A country may not need every frontier model trained entirely inside its borders, but it will increasingly ask that sensitive inference happens under local law, on trusted infrastructure, with auditable model behaviour and operational continuity. This is why AI sovereignty will manifest as a network of national and regional AI facilities, not merely one national supercomputer.

The IEA has warned that global data centre electricity consumption is projected to more than double to around 945 TWh by 2030, “slightly more than Japan’s total electricity consumption today”. [4] This turns data centre strategy into energy strategy. The best locations will not simply be near financial districts or existing cloud hubs. They will be near power, water, fibre, subsea cable routes, heat reuse opportunities, skilled labour and politically acceptable land.

The new design principles

Sovereign AI will push data centre design in five directions. First, resilience by geography. Nations will not want a single point of compute failure. They will favour dispersed clusters: primary AI factories, regional inference nodes, edge compute in telecom networks, and fall-back capacity in allied jurisdictions.

Second, energy coupling. Data centres will be planned around renewable generation, grid constraints, private wire connections, battery storage, modular power and demand response. The UK’s emerging AI Growth Zone model points in this direction: designated zones are intended to accelerate planning, power access and data centre build-out. [5] In practice, the winning design will look less like an isolated technology park and more like an integrated energy-digital-industrial campus.

Third, sovereign operations. The debate will move from “where is the building?” to “who operates it, who has administrative access, which law applies, what foreign dependencies exist, and what happens during sanctions, cyber conflict or supply chain disruption?” This explains why Europe’s recent digital sovereignty procurement has introduced assurance concepts such as Sovereignty Effectiveness Assurance Levels, with requirements around legal, operational and supply-chain resilience. [6]

Fourth, sector-specific enclaves. Banking, telecoms, logistics, healthcare and defence will not all use the same sovereignty pattern. A national model will have to support regulated partitions: confidential computing, private LLMs, verifiable audit trails, model lineage and controlled data sharing between government and industry.

Fifth, social licence. Reddit and developer forums reveal a healthy scepticism: if the chips, cables, software and cloud control planes are owned elsewhere, is sovereignty real or just branding? One Reddit contributor put the issue bluntly: “Does that count as sovereignty. Debatable. I’d say no.” [7] That sentiment matters because communities will be asked to accept more data centres, more grid reinforcement and more land use change. The value exchange has to be visible: jobs, heat reuse, skills, regional regeneration and cheaper, cleaner power.

Consequences for international companies

For telecommunications operators, sovereign AI is both a threat and an opportunity. It threatens the old carrier model where connectivity is sold as capacity alone. But it creates a new role for telcos as national AI fabric operators: providing low-latency edge compute, secure data exchange, identity, lawful intercept governance, IoT intelligence and resilience across critical networks. Operators that own fibre, mobile edge locations, data centres and trusted enterprise relationships can become anchor institutions in sovereign AI ecosystems.

For banks and financial services firms, the implications are equally material. AI models will sit inside credit, fraud, trading surveillance, customer vulnerability, cyber defence and regulatory reporting. Boards will increasingly ask: can we explain where the model ran, what data it touched, whether the regulator can audit it, and whether we can continue operating if a foreign cloud service is disrupted? The likely response is hybrid: global cloud for scale, sovereign cloud for regulated workloads, and private model environments for the most sensitive functions.

For logistics companies, AI sovereignty intersects with physical sovereignty. Ports, warehouses, shipping lanes, border systems, customs data, predictive maintenance and fleet optimisation all depend on real-time data. If AI controls routing or risk scoring, then data centre geography becomes part of supply chain resilience. A logistics group operating across Europe, Asia and the Middle East may need regional AI nodes aligned to trade corridors, not merely corporate headquarters.

For hyperscalers and international technology firms, the message is clear: sovereignty cannot be dismissed as protectionism. It is becoming a customer requirement. The winners will be those able to offer sovereign controls without destroying interoperability: local legal entities, transparent operational models, encryption and key sovereignty, local support teams, exit rights, open standards and credible partnerships with national champions.

The policy choice

There is a danger that governments confuse sovereignty with autarky. Full national independence across chips, models, cloud software, energy systems and talent is unrealistic for most countries. The better goal is strategic optionality: enough domestic and allied capability to avoid coercive dependency, enough openness to remain innovative, and enough governance to earn trust.

In practical terms, the next generation data centre will be judged not only on power usage effectiveness, cost per rack or GPU density. It will be judged on sovereign value: does it strengthen national resilience, support local industry, protect sensitive data, reduce carbon intensity, improve public services and give domestic firms a route into the AI economy?

AI sovereignty therefore marks a behavioural shift. Data centres will stop being hidden technical real estate and become visible instruments of industrial policy. Their geography will follow power and politics as much as fibre and land. Their design will embed trust, auditability and continuity. And for international companies, the strategic question will no longer be whether they use AI, but whether their AI can operate legitimately, resiliently and locally in every jurisdiction that matters.

That, I suspect, is the real meaning of AI sovereignty. It is not the ownership of a machine. It is the ability of a nation, and the companies operating within it, to shape the intelligence on which their future depends.

References and source notes

[1] OECD, “A blueprint for building national compute capacity for artificial intelligence”, 2023; OECD AI Compute topic page, 2025/26. Quote: national plans should address “security, sovereignty, sustainability”. https://www.oecd.org/en/publications/a-blueprint-for-building-national-compute-capacity-for-artificial-intelligence_876367e3-en.html and https://www.oecd.org/en/topics/ai-compute.html

[2] UK Government, AI Opportunities Action Plan: government response, 13 January 2025. Quote: “expand our sovereign compute capacity by at least 20x by 2030”. https://www.gov.uk/government/publications/ai-opportunities-action-plan-government-response/ai-opportunities-action-plan-government-response

[3] European Commission, AI Continent Action Plan factpage, 7 May 2025. Quote: AI Gigafactories will be “4x more powerful than AI Factories”; the Cloud and AI Development Act aims to “Triple the EU’s data centre capacity in the next 5-7 years”. https://digital-strategy.ec.europa.eu/en/factpages/ai-continent-action-plan

[4] International Energy Agency, Energy and AI: Executive summary, 2025. Quote: data centre electricity consumption is set to more than double to around “945 TWh by 2030”. https://www.iea.org/reports/energy-and-ai/executive-summary

[5] UK Government, AI Opportunities Action Plan: One Year On, 29 January 2026. Quote: the UK has “designated 5 AI Growth Zones, unlocking investment and accelerating data centre buildout”. https://www.gov.uk/government/publications/ai-opportunities-action-plan-one-year-on/ai-opportunities-action-plan-one-year-on

[6] ITPro, “European Commission awards digital sovereignty contracts”, 20 April 2026. Reported reference to Sovereignty Effectiveness Assurance Levels (SEAL) and digital sovereignty criteria. https://www.itpro.com/cloud/cloud-computing/european-commission-awards-digital-sovereignty-contracts-backs-google-cloud-involvement

[7] Reddit discussion on sovereign AI and infrastructure ownership, 2026. Quote: “Does that count as sovereignty. Debatable. I’d say no.” https://www.reddit.com/r/I_DONT_LIKE/comments/1r76rac/idl_how_everyones_talking_about_ai_data/

[8] NVIDIA / World Government Summit coverage, 2024. Quote attributed to Jensen Huang: “Every country needs to own the production of their own intelligence.” https://www.financemiddleeast.com/fintech/every-country-needs-sovereign-ai-says-nvidias-huang/