When Nvidia moves, the global AI ecosystem shifts. The world’s most valuable chipmaker has made yet another aggressive push into AI infrastructure, this time with a £500 million ($683 million) equity investment in London-based startup Nscale.
The deal is part of a broader £11 billion ($15 billion) commitment with Nscale and U.S. rival CoreWeave to expand the U.K.’s AI capacity — and it underscores how Nvidia is rapidly extending its influence well beyond silicon.
Nvidia’s strategy has been clear for months: invest directly in companies building the backbone of AI. The chipmaker recently took a 4% stake in Intel, and it has poured billions into CoreWeave, one of its biggest infrastructure customers. With Nscale, Nvidia is now anchoring itself to Europe’s rising push for “sovereign AI,” a bid to keep critical workloads and data on local soil.
Register for Tekedia Mini-MBA edition 19 (Feb 9 – May 2, 2026): big discounts for early bird.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register for Tekedia AI Lab: From Technical Design to Deployment (next edition begins Jan 24 2026).
Nscale’s Meteoric Rise
Just two years ago, Nscale didn’t exist as a standalone entity. It was spun out of Arkon Energy, a crypto-mining infrastructure provider, to meet surging demand for AI-specific data centers after OpenAI’s ChatGPT sparked the AI boom in late 2022.
Like CoreWeave in the U.S., Nscale repurposed its crypto roots into an AI infrastructure business, combining vast data center capacity, high-density power, and thousands of GPUs with software layers for clients.
The transformation has been astonishingly fast. In May 2023, Nscale came out of stealth, quickly partnering with the UAE’s Open Innovation AI to deploy 30,000 GPUs. It acquired Finnish high-performance computing firm Kontena, then struck a deal with Asian telecom giant Singtel to offer GPU-as-a-Service in Europe and Southeast Asia. Initially reliant on AMD hardware, Nscale pivoted to Nvidia GPUs as its partnerships deepened.
By December, it closed one of the largest Series A rounds in U.K. history — $155 million led by Sandton Capital Partners. With that, the company expanded its pipeline from 300 megawatts to 1.3 gigawatts in just one year and set a target of running 350,000 GPUs by 2027.
Microsoft and OpenAI, Too
Tuesday’s announcements signaled Nscale’s leap into the big leagues. The company was named an AI infrastructure partner for Nvidia, Microsoft, and OpenAI. It also revealed a five-year, $6.2 billion deal with Microsoft and Norwegian industrial group Aker to develop “hyperscale AI infrastructure” across Europe, with Norway as the anchor.
Microsoft committed $15.5 billion of new investment in the U.K., headlining plans to build the country’s largest supercomputer in Loughton, Essex. That site will house 23,040 Nvidia Blackwell GPUs by early 2027, providing 50 megawatts of AI power, scalable to 90 megawatts.
“No one can make that kind of capital investment unless they’ve got somebody already committed to spend the money once the work is complete, and that’s the role we’re playing,” Microsoft President Brad Smith explained.
Meanwhile, OpenAI confirmed it would launch Stargate U.K. in partnership with Nscale and Nvidia, initially deploying 8,000 GPUs next year, with the capacity to expand to 31,000. The sites will span several regions, beginning with Newcastle’s Cobalt Park.
Nvidia’s CEO Jensen Huang personally announced the £500 million equity injection, telling reporters: “We convinced ourselves that Nscale could be a national champion for AI infrastructure in the U.K.”
Europe’s Sovereign AI Gamble
Nscale’s emergence coincides with Europe’s drive for “sovereign AI” — a policy push requiring data centers and AI workloads to remain within European jurisdictions. Britain’s AI action plan, rolled out in January, promised to slash bureaucratic barriers to domestic AI growth.
Josh Payne, Nscale’s founder and CEO, framed the Microsoft and Aker partnership as a “huge win for European-owned AI infrastructure.” In earlier remarks, he said Europe’s biggest challenges were a lack of computing capacity and a “very fragmented market.” His vision is to create massive projects that smaller players can consume from, unlocking productivity and economic growth.
Nick Patience, AI practice lead at Futurum Group, said the announcements signal government recognition that “it has to do something to get the AI infrastructure built here, which has been a long slog.”
Scaling Targets, Heavy Costs
Payne has set ambitious targets: 50,000 GPUs by the end of 2025 and 150,000 by the end of 2026. Long term, Nscale aims to reach 350,000 GPUs by 2027. Its Norwegian project with OpenAI alone has a $1 billion commitment, with plans to host 100,000 GPUs before 2027.
But scaling at this pace is intensely capital-intensive. Building AI data centers requires billions in both debt and equity. CoreWeave has already raised $12.4 billion in debt and over $1 billion in equity, plus a $1.5 billion bond issue in July after a $2 billion debt sale in May.
Nscale is following a similar path, reported earlier this year to be raising $1.8 billion in private credit through Goldman Sachs.
Outlook
Analysts sketch three possible paths for the partnerships. In the optimistic case, Nscale leverages its Microsoft, Nvidia, and OpenAI alliances to secure long-term contracts and dominate Europe’s AI infrastructure landscape, emerging as a “CoreWeave of Europe.” In the base case, it achieves steady growth but faces periodic financing strains, forcing slower buildouts. The downside scenario sees mounting debt, regulatory bottlenecks, or execution delays stalling projects, leaving Nscale vulnerable to acquisition.