Nscale's £2 billion AI data center faces delays due to the UK power grid.
Britain aims to become a leader in AI. However, one of its major data centres is unable to obtain the necessary electricity to operate. Nscale's £2bn facility in Essex has the funding, planning approval, and a grid connection, but is lacking power to start up on schedule.
Nscale possesses the finances, the location, and a major client, but it does not have a power supply. The British AI company, supported by Nvidia, is constructing a £2bn data centre in Loughton, Essex, with Microsoft as its primary tenant. After securing planning approval and a 90-megawatt grid connection, the company found out that the power would not be available in time for the planned opening next year, as reported by Sifted.
In response, the company is seeking alternative electricity solutions. It is exploring on-site power generation options, including discussions with a California fuel-cell manufacturer, instead of depending on the National Grid. A spokesperson confirmed their ongoing commitment to the Loughton project.
The issue Nscale faces is indicative of a broader problem in Britain. Delays in grid connections have become a significant obstacle for new data centres, with some UK projects experiencing waits of nearly a decade for sufficient power. An estimate indicated that over 25% of data-centre capacity was delayed in 2025 alone.
Consequently, many projects are considering off-grid solutions. According to industry data, more than 100 UK initiatives are currently evaluating gas generation while awaiting grid connections. This poses a problem for a government aiming to electrify the economy and enhance the cleanliness of its power sources simultaneously.
The political landscape is challenging. Sir Keir Starmer’s administration has prioritized data centres in its growth strategy, designating them as essential national infrastructure. Unfortunately, the grid has not evolved alongside this ambition. Taco Engelaar from infrastructure firm Neara cautioned that “Britain’s aspirations for data centre independence will vanish if we fail to address grid issues.” A recent forecast suggests that the nation may fall short of its 2030 clean-power goal, achieving around 83% instead of the targeted 95%.
Despite these challenges, Nscale remains confident, financially robust, and dynamic. It increased its workforce by 121% in the first half of 2026, making it one of Europe’s fastest-growing companies, and secured a $900m credit facility just weeks ago. The constraints are not financial but rather physical.
The delay in Essex is a reflection of a larger trend occurring throughout the West. New York has recently become the first U.S. state to impose a freeze on new data centres, while Scotland’s governing party is advocating for a similar pause. Meanwhile, the surge in data centres has triggered the most extensive gas-plant construction boom in history in the U.S. Researchers warn that Europe’s aspirations for sovereign AI might be similarly hindered by these limitations.
The message is clear yet unwelcoming: AI is marketed as software, but it fundamentally relies on electricity, which, unlike capital, cannot be mobilized overnight.
Other articles
Hassabis seeks a regulatory body similar to FINRA for advanced AI.
Demis Hassabis advocates for the establishment of a US-led AI Standards Body, inspired by Wall Street’s FINRA, to evaluate frontier models prior to their launch and to potentially decelerate the industry when necessary.
Medical AI has supplanted nurses and evaded its own regulatory checks.
Two cases in the US highlight the negative aspects of medical AI: nurses at Montefiore were substituted by software, and a whistleblower from Mayo Clinic claims that a tool has a 67% error rate.
Medical AI substituted for nurses and evaded its own regulations.
Two cases in the US highlight the negative aspects of medical AI: nurses at Montefiore being replaced by software, and a whistleblower from the Mayo Clinic claiming that a tool has a 67% error rate.
Kalshi develops a forward curve for computing power as exchanges compete to transform GPUs into a tradable asset.
The prediction market platform Kalshi has developed a forward curve to monitor GPU compute expenses, joining CME and ICE in the effort to monetize AI infrastructure.
X is instructing its AI algorithm on concepts that social networks used to grasp.
X is enhancing posts among mutual followers after discovering that its algorithm lacked relationship data, highlighting the discrepancy between forecasting engagement and fostering discussions that people truly desire.
The vision of Concorde is attainable. NASA's X-59 simply needs to eliminate the sonic boom from supersonic flight.
NASA's X-59 has achieved speeds of 924 mph at an altitude of 55,000 feet, moving the agency closer to demonstrating that future supersonic passenger flights can circumvent loud sonic booms.
Nscale's £2 billion AI data center faces delays due to the UK power grid.
Nscale's £2bn data centre in Essex is unable to secure grid power in time for its launch, leading the company to pursue its own electricity sources as connection wait times in the UK extend for years.