OpenAI pauses UK data centre deal over energy costs and regulation
openai regulation
| Source: BBC on MSN | Original article
OpenAI has put its £2 billion “Stargate UK” data‑centre project on hold, citing soaring energy prices and an unfavourable regulatory climate. The initiative, a joint effort with Nvidia and the UK‑based cloud provider Nscale, was slated to install up to 8 000 GPUs initially, with a longer‑term vision of scaling to 31 000 units. The pause was announced in a brief statement to Reuters, which added that the company will continue to explore the venture when “the right conditions enable long‑term infrastructure investment.”
The development strikes a blow to the UK government’s ambition to brand the country as an AI superpower. Earlier this month, the administration bundled the data‑centre plan into a broader tech‑investment package that promised thousands of high‑skill jobs and a competitive edge in generative‑AI research. As we reported on 9 April, OpenAI had already shelved a £31 billion UK investment programme amid fiscal and policy concerns; the current suspension deepens that setback.
Energy costs matter because AI training workloads are among the most power‑hungry commercial applications. Britain’s recent carbon‑pricing reforms and the push for net‑zero have driven electricity tariffs higher than in many rival locations, eroding the economic case for large‑scale compute clusters. At the same time, regulators are tightening data‑centre licensing and safety standards, adding uncertainty for foreign investors.
What to watch next includes a possible policy response from the Department for Business and Trade, which may tweak incentives or streamline approvals to retain AI capital. Analysts will also monitor whether OpenAI shifts its compute strategy toward other European sites or accelerates its own renewable‑energy projects. Finally, the pause could ripple through the UK’s broader AI ecosystem, influencing the timing of related ventures from DeepMind, Graphcore and other home‑grown players seeking to ride the generative‑AI wave.
Sources
Back to AIPULSEN