(O-IAHI/Shutterstock)
Nvidia has announced a pair of deals this week that highlight both its commercial cloud strategy and its role in national AI initiatives. On Monday, CoreWeave, the GPU cloud specialist backed by Nvidia, announced a $6.3 billion agreement that effectively guarantees cloud capacity through the next decade, with Nvidia itself stepping in to purchase any unused resources. On Tuesday, CEO Jensen Huang arrived in London to announce plans for the largest AI infrastructure rollout in U.K. history, part of a push to make Britain a sovereign AI power.
(PJ McDonnell/Shutterstock)
The two announcements reflect different but complementary priorities for the company. The $6.3 billion CoreWeave contract builds on a prior deal from April 2023 and extends through 2032. Under its terms, Nvidia will purchase any unused CoreWeave capacity, ideally giving the datacenter operator a decade of financial certainty as it continues to grow. CoreWeave has emerged as a prominent provider of GPU cloud services and operating facilities in the U.S. and Europe, supplying compute for companies including OpenAI and Microsoft. In March, CoreWeave and OpenAI signed a five-year deal worth $11.9 billion, followed by another agreement committing up to $4 billion in payments through 2029.
Nvidia’s commitments in the U.K. are on a larger scale and are aligned with the nation’s goal of sovereign AI capacity. The company and its partners plan to invest up to £11 billion in what Nvidia calls AI factories, or large datacenters built around Nvidia’s new Blackwell Ultra GPUs. By 2026, as many as 120,000 of these chips are expected to be operating in Britain, making it the largest AI infrastructure rollout in the nation’s history, according to Nvidia.
The initiative is part of a collaboration first announced by Prime Minister Keir Starmer and Nvidia CEO Jensen Huang in June. In a panel with Starmer at Tech Week London, Huang described the U.K. as a “Goldilocks” environment for AI, citing the country’s mix of expertise, infrastructure and emerging opportunities as reasons for new investment. Now, the goal is framed as a way to make the U.K. an “AI maker, not an AI taker,” with locally hosted compute for projects ranging from national language models to advanced medical research. Nvidia is working with Nscale, OpenAI, Microsoft and CoreWeave to build and operate the facilities, including Stargate U.K., a planned system expected to serve OpenAI’s latest models.
(Source: Nvidia)
Alongside the hardware investments, Nvidia outlined research collaborations across multiple fields. Projects include a large language model designed to support both English and Welsh, a multimodal health model called Nightingale AI, a high-resolution pollution model, and new foundation models for medical imaging and materials science. Robotics firms such as Extend Robotics and Oxa are also adopting Nvidia platforms for manufacturing and autonomous driving, while life sciences companies including Isomorphic Labs and Oxford Nanopore are using Nvidia tools for drug discovery and genomics.
These developments show how access to compute power has become a key factor in the global adoption of AI. With similar sovereign AI projects underway in Europe and Asia, the U.K. rollout may serve as a template for future national strategies. For now, Nvidia’s dual announcements show how it is using its dominant hardware position to secure a significant role in the infrastructure and policy decisions driving the AI economy.