Bold reality check: Britain’s AI ambitions may stall if the energy backing it stalls first.
Britain’s AI surge is running straight into an energy bottleneck, with nuclear power poised to be the dependable backbone. Yet the government has paused progress just as demand from data centers is about to surge, leaving investors wondering if the UK will talk itself out of its AI opportunity.
Recent analysis from the Nuclear Industry Association and Oxford Economics projects data-center electricity demand to rise more than fivefold by 2030, consuming nearly 9% of the UK’s total electricity. The AI labs and hyperscalers driving this increase want plug-in-ready, 24/7 power availability within two years, but the current grid-connection timeline in the UK stretches to about ten years.
This backdrop helps explain Rachel Reeves’ decision to pause a sweeping package of planning reforms that had aimed to streamline nuclear development. A consultancy review, led by Fingleton, popularized the term ‘fish disco’ as a symbol of regulatory overreach. The intent was to clear regulatory undergrowth, but concerns raised by a single adviser triggered a halt for “further work.”
Industry voices emphasize that nuclear projects are inherently challenging—long timelines and high costs require consistent, predictable policy. George Borovas, head of Habton Andrews Kurth’s global nuclear practice, told City AM that unpredictability from government creates investor nerves. He warned that this moment is precisely when reliability matters most: investors want smooth, predictable progress.
Ironically, the timing is off. A few years ago, hyperscalers prioritized renewables; now they’re turning to nuclear as a must-have, not a nice-to-have. If a country cannot guarantee 24/7, predictable power, AI firms will relocate elsewhere.
On a brighter note, the government has chosen Wylfa in Anglesey to host the UK’s first three Rolls-Royce small modular reactors (SMRs), designed to deliver more stable, low-carbon power. After years of false starts and Hitachi’s large-scale write-off, this marks a constructive step that could anchor thousands of UK jobs and nurture a domestic SMR supply chain. Yet SMRs are not a silver bullet. Borovas cautions that the real challenge lies in assembling project structures that allow rapid replication and fleet-building; individual projects remain difficult, and investors may be wary of a single-project pitch.
The UK’s financing framework for nuclear is a genuine strength. Borovas notes that the CFD model and risk-sharing approaches resemble Thames Tideway’s model, placing the UK at a leading edge. That asset makes today’s pause especially consequential: the country has built a respected framework and now hesitates on reforms needed to scale.
If Britain wants to lead in the AI race, nuclear must accelerate to match demand. Clear planning, shorter timelines, and decisive regulation are essential to translate energy policy into AI infrastructure. Without energy certainty, AI firms will relocate. And as Borovas says, many already seem to be watching closely and weighing their options, waiting to see if Britain can deliver.
Thought-provoking question for readers: Do you think the UK’s current approach to nuclear reform is enough to unlock AI-scale energy demand, or should policymakers prioritize rapid, pragmatic solutions even if they mean taking bolder risks now? Share your view in the comments.
