OpenAI halts Stargate UK due to energy expenses and copyright regulations obstructing progress.

      In summary: OpenAI has suspended its Stargate UK data centre initiative, citing high industrial electricity costs in Britain and an adverse regulatory climate concerning AI copyright. The project, which was unveiled in September 2025 in collaboration with Nvidia and Nscale, aimed to establish 8,000 GPUs in north-east England, with plans to expand to 31,000 over time. OpenAI indicated it would proceed “when the right conditions” arise, but has not provided a specific timeline. This pause represents a notable setback for the UK government's AI Growth Zones initiative and coincides with OpenAI's preparations for a public listing.

      What Stargate UK was intended to be

      Stargate UK was introduced in September 2025 as a sovereign AI infrastructure collaboration among OpenAI, Nvidia, and UK cloud service provider Nscale, designed to enhance data centre capabilities in north-east England, allowing OpenAI's models to operate on local computing resources. The proposed sites included Cobalt Park near Newcastle and Blyth, both located within the UK government's established AI Growth Zones, which the government deemed central to its industrial AI strategy. The project was revealed during US President Donald Trump’s state visit to the UK, adding diplomatic significance to its commercial aspects. The initial phase involved acquiring about 8,000 Nvidia AI processors, with aspirations to scale that number to 31,000 GPUs over time. This capacity was intended to enable OpenAI to support vital public services, regulated sectors like finance, and national security collaborations without transferring data through US-based infrastructure. OpenAI has not publicly shared the total investment for the UK project. Meanwhile, the broader US Stargate initiative continues to progress, with data centre construction underway throughout the United States, supported by a $40 billion bridge loan from SoftBank to finance its involvement, indicating that the UK pause is an isolated geographical exception rather than a withdrawal from overall AI infrastructure investment.

      The issue of energy costs

      The primary challenge identified by OpenAI is the high cost of electricity in the UK. Industrial electricity prices in the UK rank among the highest of any IEA member country, exceeding four times those in the United States, Finland, Norway, and Sweden. For a data centre consuming 100 megawatts, this pricing discrepancy poses a structural issue rather than just a line-item concern: the economic viability of running large-scale AI inference tasks in an area where power costs are considerably higher than in Virginia or Texas alters significantly, and the gap becomes more pronounced as capacity increases. This challenge goes beyond merely electricity rates. The demand for grid connections in the UK soared from 41 gigawatts in November 2024 to 125 gigawatts by June 2025, with an estimated 75 gigawatts of that total related to data centre initiatives. While buildings can be established in 18 to 24 months, securing grid connections takes between three to eight years. This discrepancy indicates that even if a project overcomes the financial challenge, it still faces an infrastructure backlog that the current regulatory and planning systems are not equipped to handle at the speed required for AI infrastructure. The UK government's AI Growth Zones policy, released in November 2025, aimed to address this bottleneck, but the designated zones do not resolve the fundamental grid limitations, and OpenAI’s decision to pause indicates that the policy hasn't yet created the necessary conditions for investment viability.

      The copyright issue

      Alongside energy costs, OpenAI highlighted a regulatory concern regarding the UK's ambiguous approach to AI copyright. UK legislators have been attempting to revise the regulations regarding the training of AI models on copyrighted content. The government's preferred method, which suggested a broad text and data mining exemption with an opt-out option for rights holders, was largely rejected by respondents to its consultation, with creative sectors, publishers, and media organizations arguing that such an exemption would enable generative AI companies to train on their works without fair compensation or adequate consent. The consultation yielded no consensus, and legislative changes have since been postponed. For OpenAI, which trains large language models on internet-sourced text, the uncertainty surrounding the legality and terms of that training poses a significant business risk. A UK data centre is not merely a power facility; it also establishes legal jurisdiction. If the UK ultimately adopts a copyright framework that imposes stricter limits on training data use than the US, operating in Britain could subject OpenAI to liabilities or compliance costs not applicable to its US activities. The pause provides OpenAI the opportunity to await clarity on this regulatory landscape before investing capital.

      A pause, not a cancellation, with IPO considerations

      OpenAI’s statement was crafted to indicate that it remains open to future possibilities. “We continue to explore Stargate UK and will move forward when the right conditions, such as regulation and energy costs, enable long-term infrastructure investment,” the company stated, portraying the decision as provisional rather than definitive. However, the timing is noteworthy. OpenAI completed a $122 billion funding round at an $852 billion valuation in late March 2026, allowing