The UK talks a good game on artificial intelligence (AI), with ministers championing AI Growth Zones and a UK Sovereign AI fund, but behind the scenes Britain’s ambitions are colliding with an unglamorous but critical bottleneck. Hyperscale data centres, which are essential for training and deploying advanced AI learning models, require vast amounts of power and land. Despite this, developers face long waits for grid connections, high energy prices, and a planning system seemingly ill equipped for the scale and urgency of this demand.
The most striking example came this month when OpenAI paused its flagship Stargate UK data centre project, citing all of the above. What was meant to be a cornerstone of Britain’s AI strategy has been pushed into the long grass.
This is a big problem, and companies are starting to look elsewhere. The US continues to surge ahead with aggressive investment and faster approvals, while parts of Europe are also streamlining processes to attract new operators.
AI capability increasingly hinges on access to computing power and without it UK firms may become increasingly dependent on overseas infrastructure, raising costs and eroding sovereignty in a technology set to shape everything from finance to defence. Data centres already make up around 2 percent of Britain’s electricity demand and this is expected to more than double by 2030.
For the government, the fix will not be an easy one. The UK continues to have major strengths with world-leading research capabilities, a buoyant start-up environment and deep capital markets, but to prosper in AI computing power is a non-negotiable, and at the moment it seems that Britain is stuck in the queue.
Capital at risk. All views expressed are those of the author and should not be considered a recommendation or solicitation to buy or sell any products or securities.
