Has the UK’s AI infrastructure development been a success?

When the UK announced its AI Opportunity Action Plan in January – a grand blueprint for rolling out the technology across society – Prime Minister Keir Starmer declared the strategy would make the country an “AI superpower”.

One of the key pillars of this plan was the rapid construction of data centers that could provide the massive computing requirements needed to deploy AI. This will be driven by “AI Growth Zones” – designated areas with relaxed planning permission and improved access to electricity.

Almost a year has passed since then, Nvidia, microsoftand google Each country is pumping billions of dollars into domestic AI infrastructure. Four AI growth zones have been identified, with homegrown startups like Nscale emerging as key players in the space.

But critics point to severely limited access to energy through the national grid and construction delays as signs the country is at risk of falling further behind global rivals in the AI ​​race.

Ben Pritchard, CEO of data center power provider AVK, told CNBC that “ambition and reality are still not aligned.”

“Growth has been largely held back by constraints around electricity availability, with grid bottlenecks in particular slowing the pace of development and meaning the UK is not yet deploying infrastructure fast enough to keep pace with global competitors.”

AI Growth Zones are currently in the early stages of development, so the UK’s AI infrastructure build-out is still in its infancy.

Construction work has not yet begun on the Oxfordshire site, which was first announced in February, and the company is considering proposals for delivery partners. It was announced in September that ground clearing work has begun on the facility in north-east England, with formal construction expected to begin in early 2026.

Two further locations in North Wales and South Wales were unveiled in November. The former is looking for an investment partner, and the Department of Science, Technology and Innovation (DSIT) told CNBC that a partner is expected to be confirmed in the coming months. The latter consists of a series of sites, some of which are already operational and additional construction work will be carried out on others, DSIT said.

In July, the UK government announced that it was targeting a core group of AI growth zones to serve at least 500 megawatts of demand by 2030, with at least one growing to more than 1 gigawatt by then.

But Mr Pritchard said the most serious challenge to realizing these ambitions was the UK’s limited grid capacity.

“Developers are anticipating grid connection delays of eight to 10 years, and the amount of outstanding connection requests is unprecedented, especially around London,” he told CNBC.

As businesses and consumers begin to use AI, AI workloads are also “dramatically increasing energy demands,” Pritchard added, putting further pressure on strained energy systems. “They are no longer isolated risks, but are actively slowing or preventing development across countries.”

According to Kao Data’s Spencer Lam, the public call for applications for the AI ​​Growth Zone concept has created a situation in which landowners with steel towers or power cables on their property can apply for designation.

“As a result, the national grid was inundated with grid applications from speculators,” he told CNBC, adding that there was no realistic chance of success.

The National Energy System Operator (Neso), the UK public body responsible for managing the national electricity grid, has moved to resolve the situation.

Earlier this month, the company announced plans to prioritize hundreds of projects to speed up access to the power grid. In response to a question from CNBC, Neso declined to comment on whether AI infrastructure projects would be among the priority projects, but said a significant portion would be data centers.

There has also been huge funding from tech giants, much of which was touted by the UK government in September.

Microsoft, Nvidia, Google, OpenAI, CoreWeave and others announced billions of dollars in AI investments during President Donald Trump’s state visit, including plans to bring the latest chips and open new data centers in the country.

Nscale, a homegrown startup providing access to AI computing and building data centres, also announced a deal to deploy tens of thousands of Nvidia chips at its AI factory outside London by early 2027.

“Investment from large private companies has laid an important foundation,” Puneet Gupta, general manager for the UK and Ireland at data infrastructure company NetApp, told CNBC. “Momentum is also building around plans for the country’s research supercomputers and new computing capacity, including a promise to build an AI ‘gigafactory’ in the UK.”

But the “real test” will be how quickly these plans are translated into usable computing for UK organizations, said Mr Gupta.

Stuart Abbott, managing director for the UK and Ireland at AI infrastructure firm VAST Data, told CNBC that long-term success in building the country’s AI infrastructure will require investment in the “full stack”, including data pipelines, storage, energy procurement, security, people and skills.

That means “developing operational structures that allow real-world agencies to safely deploy AI at scale,” he added. “If the UK wants this to be sustainable and not just a year-long sugar rush, we need to treat AI infrastructure like economic infrastructure.”

The challenges are significant. The value of European data center deals pales in comparison to the amount of money pumped into US projects. The UK also currently has the most expensive energy in Europe, around 75% more expensive than it was before Russia’s invasion of Ukraine, as well as legacy grid infrastructure that could take years to connect to new sites.

AVK’s Pritchard said one potential solution for projects that cannot secure access to the national grid is microgrids. A microgrid is a self-contained power network powered by engines, renewable energy, batteries, and other power sources.

AVK is currently designing two microgrids in the UK, not for AI, but for partners building cloud computing. Pritchard said it would take about three years to build and currently costs about 10% more than energy from the grid.

VAST Data’s Abbott said co-locating computing where power already exists, rather than “forcing everything to be greenfield” (a term used to refer to greenfield sites), is also a way to get AI infrastructure up and running faster.

Kao Data’s Lam told CNBC that the pace of implementation will be important. “Unless fundamental issues around energy availability and prices, AI copyrights and funding for AI development are resolved quickly, the UK risks missing out on one of the most remarkable economic opportunities of our time and ultimately becoming an international AI backwater.”