Nvidia and United Kingdom Build Nation's AI Infrastructure (nvidianews.nvidia.com)

NVIDIA announced a major buildout of U.K. AI infrastructure with partners including Nscale, CoreWeave, Microsoft and OpenAI: up to 120,000 NVIDIA Blackwell GPUs and as much as £11 billion of local investment — the largest AI infrastructure rollout in U.K. history — plus Nscale’s wider deployment of 300,000 NVIDIA Grace Blackwell GPUs worldwide (60,000 in the U.K.). By the end of 2026 the program will host “AI factories” to serve and train large models (including OpenAI’s Stargate U.K. and expected GPT-5 workloads), a Microsoft-backed supercomputer in Loughton with >24,000 Grace Blackwell Ultra GPUs, CoreWeave’s renewable-powered data center in Scotland, and BlackRock-funded data center modernization to be NVIDIA-ready. Technically this tightens compute, data‑sovereignty and low‑latency hosting for national research and industry: Blackwell and Grace‑series GPUs, DGX Cloud access, NVIDIA AI Technology Centers, and upskilling via NVIDIA Deep Learning Institute will accelerate model development, inference and generative/agentic applications across medicine, robotics and industry. NVIDIA is also linking GPUs and quantum tech — CUDA‑Q integrations with OQC, hybrid photonic QPUs, and GPU‑accelerated error‑correction work at universities — signaling practical hybrid quantum–GPU workflows. For the AI/ML community this means far greater local capacity for training and serving state‑of‑the‑art models, enhanced collaboration between academia and industry, and a national platform aimed at sovereign, sustainable AI innovation.