There is a flattering story doing the rounds in Westminster, and it is half right. Brexit, the argument goes, spared Britain the European Union’s safety-first AI Act, leaving the country free to set a lighter, more pragmatic rulebook and steal a march on a continent that chose paperwork over progress. The flattering part is true: the UK does have regulatory room that the EU’s 27 member states do not. The part that gets left out is harder. A lighter rulebook is only an advantage if you have the compute and the energy to act on it, and on both counts Britain is behind. The regulatory gap is a window, not a moat, and windows close.
The story Brussels told itself
When the European Parliament reached agreement on the AI Act late in 2023, EU officials celebrated as though they had won the race rather than written the rules for it. The images from Brussels — lawmakers applauding a legal framework — landed awkwardly next to what the rest of the world was shipping. The Silicon Valley shorthand wrote itself: “America builds, China scales, Europe regulates.” The Act came into force in August 2024, and the self-congratulation resumed.
The problem was never that the EU regulated. Rules are not the enemy of a serious AI economy, and several of the Act’s concerns — data protection, workers’ rights, transparency — are legitimate. The problem was that Brussels treated regulation as the main event. While the EU debated guardrails, the United States and China were assembling the things that actually decide who leads: chip manufacturing, compute capacity, energy supply and frontier labs. By the time the Act was law, the contest had moved to ground it barely addressed.
Strategic Reality: Regulation governs how AI is used. It does not produce the compute, the energy or the capital that determine who builds the frontier in the first place. Mistaking the first for the second is how you legislate yourself out of a race you think you are refereeing.
This is the context for the Europe 2031 scenario published last week by a group of European researchers and academics — a deliberately bleak forecast of a continent first left behind and then squeezed into irrelevance, eventually cut off from US-grown AI by executive order in 2029. The unsettling detail is that the fiction has already started arriving early. Washington moved to restrict foreign access to a leading US lab’s newest models this month, years ahead of the scenario’s timeline. We covered what that means for Britain in Trump’s Anthropic ban exposes Britain’s AI dependence. Reality is not just catching up with the forecast; it is running ahead of it.
Where Britain actually sits
The honest picture is more interesting than either the triumphalism or the despair. Britain has genuine structural advantages, and genuine, expensive gaps. Holding both in view at once is the whole point.
| Lever | EU | UK |
|---|---|---|
| Regulatory freedom | Bound to the AI Act across 27 states | Free to set its own, lighter framework |
| Speed of rule-making | Consensus across member states; slow | Single jurisdiction; can move fast |
| Frontier alliance access | Tier below the US’s closest partners | Top cohort via Five Eyes |
| Compute capacity | Limited relative to US and China | Limited; behind where it needs to be |
| Energy for AI infrastructure | Constrained and costly | Among the higher industrial energy costs in the G7 |
The left two columns are where the flattering story lives. The UK can write proportionate rules without negotiating across a bloc, and it sits in a better position than most of Europe when access to frontier models is rationed by alliance. The right-hand column is where the story falls quiet. Regulatory freedom is the permission to compete. Compute and energy are the ability to. Britain has more of the first than the second, and the second is the one that costs billions and takes years.
Critical Context: A faster rulebook is a real edge over the EU. But the binding constraint on UK AI is not the speed of its regulation — it is the cost of its electricity and the scarcity of its compute. Solving the easy problem does not relieve the hard one.
The advantage that has to be spent, not banked
The mistake to avoid is treating regulatory divergence as a position to defend rather than a head start to use. A lighter rulebook earns nothing on its own. It only pays out if firms, investors and the state convert the freedom into things that are hard to copy: data centres, trained models, energy contracts, talent that chooses to stay.
That conversion is not automatic, and the clock is unfriendly. The EU can change its mind; a future UK government can tighten rules; energy and compute take years to build regardless of who is in office. The window where Britain has both regulatory room and a credible shot at the inputs is open now and will not stay open by default.
Competitive Reality: Light-touch regulation is a magnet for AI investment only if the power and compute are there to host it. Without them, a friendly rulebook just points capital toward somewhere it can actually plug in.
For business leaders, the practical reading is that the UK’s regulatory advantage is something your strategy can lean on for the next few years, but not something to assume is permanent. It is a reason to build here now, whilst the gap is real, rather than a guarantee that building here will stay easy.
Who this lands on
| Stakeholder | What the regulatory gap means in practice |
|---|---|
| UK AI startups | Faster path to market than EU peers, if compute can be sourced affordably |
| Firms scaling AI | Room to deploy without the AI Act’s compliance overhead; energy cost still bites |
| Investors | A more permissive jurisdiction to back, provided infrastructure keeps pace |
| Public sector | Freedom to adopt quickly, exposure to the same compute and supply constraints |
| EU-exposed UK firms | Must still meet the AI Act when selling into Europe, regardless of UK rules |
That last row matters and is easily missed. Regulatory divergence does not exempt a UK company from the AI Act when it sells into the single market. Britain’s lighter rules govern what you may build and deploy at home; they do nothing to lift the obligations on what you ship across the Channel.
What to do with a window
The strategy that fits a closing window is different from the one that fits a durable moat. You move whilst you can, and you build the things that outlast the advantage that let you start.
For organisations early in adoption, treat the UK’s lighter regime as a reason to move now rather than wait for certainty. The compliance cost of deploying AI here is genuinely lower than across the EU, and that is a real, present advantage. Use it to ship and learn while competitors on the continent are still mapping their AI Act obligations.
For organisations scaling AI, get ahead of the binding constraint, not the visible one. The risk to your roadmap is far more likely to be the price and availability of compute and power than a regulatory clampdown. Lock in capacity and energy terms where you can, and design so that a single squeeze on either does not stall the whole programme.
For organisations where AI is core to the business, treat the regulatory gap as a strategic timing question. The case for building UK AI infrastructure — or anchoring your operations to those who are — is strongest while the divergence holds and weakest once it narrows. Decisions that depend on the window being open should be made on the assumption that it is finite.
Take Action: Ask which of your AI plans depend on the UK’s lighter regulation, and which depend on cheap, available compute and power. If the honest answer is that you have optimised for the rulebook and ignored the inputs, you have prepared for the wrong constraint.
The challenges leaders tend to miss
Four difficulties hide behind the comfortable version of this story, and each has a practical response.
The first is mistaking permission for capability. A lighter rulebook tells you what you are allowed to build; it says nothing about whether you can afford the compute to build it. The fix is to plan around the input you actually lack — power and processing — rather than the constraint you have already escaped.
The second is assuming the divergence is permanent. The EU could soften the AI Act under competitive pressure, and a future UK government could tighten its own rules after the first serious AI harm. Build a strategy that still works if the regulatory gap narrows, because betting the business on a policy gap is betting on politics.
The third is the energy blind spot. Large-scale AI runs on electricity, and Britain’s industrial power costs are among the highest in the G7. A firm can win on every other dimension and still find that the economics of running serious AI infrastructure at home do not close. Factor real energy costs into any plan that involves training or heavy inference on UK soil.
The fourth is the export trap. A UK company that builds under light domestic rules but sells into Europe still answers to the AI Act. Treating the UK’s regime as if it travels with your product is how compliance surprises happen at the worst moment — when you are trying to expand.
Warning ⚠️: “We are lightly regulated at home” is not a market-access strategy. The moment you sell into the EU, the AI Act applies in full, and a domestic advantage becomes an export liability if you have not planned for it.
The strategic takeaway
The City AM argument that Europe has regulated itself out of the race is largely fair, and the warning that Britain has “a huge amount of catching up to do” is the more important half of the same sentence. The UK’s advantage over the EU is real. It is also unearned, in the precise sense that the country has been handed regulatory freedom by Brexit but has not yet built the compute and energy that turn freedom into capability. An advantage you have not converted is not yet an advantage; it is an opportunity with an expiry date.
For UK business leaders, three commitments separate the firms that use this moment from the ones that merely enjoy it. Move whilst the regulatory gap is open, because the lower compliance burden is a present, usable edge over EU rivals. Build or secure access to the inputs that outlast the gap — compute and energy — because those are what the contest is actually fought over. And plan for the divergence to narrow rather than assuming it will hold, so that your strategy survives the EU getting faster or Britain getting stricter.
Success Factor: The firms that come out of the next few years ahead will not be the ones that simply enjoyed a lighter rulebook. They will be the ones that spent the regulatory head start building things — capacity, models, customers — that stay valuable long after the rules on either side of the Channel converge again.
It may, as the op-ed puts it, be too late for Europe. It is not too late for Britain. But “not too late” is a statement about timing, and timing is the one advantage that disappears if you wait to be certain. The window is open now. What gets built whilst it is open is the only part of this that lasts.
Source and attribution
This analysis responds to “The EU has regulated itself out of the AI race but the UK is still in the game”, published by City AM on 16 June 2026, which sets out the post-Brexit regulatory-divergence argument and the caveat that UK compute and energy supply must catch up. It also references the Europe 2031 scenario by a coalition of European researchers and academics, and connects to our companion piece on Britain’s frontier-model dependence. The framing of regulatory freedom as a window rather than a moat, the stakeholder analysis and the recommendations are original to Resultsense.
Resultsense provides analysis to help UK professionals and businesses make sense of AI developments. For more, explore our insights and news coverage, or get in touch to discuss how these shifts affect your organisation.