OpenAI’s George Osborne Says UK Datacentre Opposition Risks Shifting AI Capacity Overseas

OpenAI’s George Osborne says resistance to new datacentres is making them hard to build in Britain—and warns that blocking them could leave the UK with less control over where the physical infrastructure behind AI is located.

Speaking at the Big Brue festival in Somerset, Osborne drew on his experience as a former politician to describe the problem as one of local consent. He said communities can resist major projects such as roads, railways, airport runways—and now datacentres. Osborne is OpenAI’s head of AI for countries, a role that includes representing the company to governments and helping enable datacentre infrastructure.

A fight over where capacity is built

Osborne’s case is not simply that Britain needs more computing capacity. He argues that if facilities cannot be built domestically, they will be built in US states including Ohio, Georgia and Michigan instead. In his framing, that would reduce Britain’s sovereign control over the physical location of its technology.

They’re going to be built elsewhere, and you know that’s a loss of some sovereign control over the physical location where these things are.

George Osborne, OpenAI’s head of AI for countries

Local objections are broader than the building

That argument meets objections beyond industrial-scale buildings. Protests have taken place in central London, Devon and outside the Scottish parliament in Edinburgh. Opponents have raised concerns about effects on landscapes and local jobs, whether facilities take space and investment that could support affordable housing, and their water and energy use.

The political disagreement is explicit. The UK government has classified AI datacentres as critical national infrastructure to help speed their construction. Meanwhile, the Green Party in England has called for a moratorium on new datacentre building. Osborne dismissed much of the opposition as nimbyism, while opponents’ reported concerns include resource use, housing and local employment.

Sovereignty has practical limits

Osborne’s warning aligns with the government’s stated goal of accelerating domestic AI datacentre capacity through AI Growth Zones. But the government’s own policy paper identifies slow, inconsistent planning and delays in gaining access to power as major barriers to investment. Local opposition is therefore one constraint within a wider buildout problem, alongside the ability to secure electricity connections.

The government does not present domestic capacity as an all-or-nothing objective: it says some AI workloads can be served offshore while Britain retains capability where it matters most. That leaves the central question unresolved: how much domestic infrastructure is necessary for resilience and control, and what conditions should communities be able to demand before accepting it?