Benzinga·2025-10-15 13:06Core Insights - Nscale has announced an expanded partnership with Microsoft to deploy approximately 200,000 NVIDIA GB300 GPUs across Europe and the United States, marking a significant AI infrastructure rollout [1][6] Deployment Details - The program will span four countries on two continents to provide hyperscale NVIDIA AI capacity for Microsoft's customers [2] - Initial phases are set to begin in 2026, with staged deliveries that will include new greenfield development and colocation footprints [3] Specific Locations and Capacities - Nscale targets around 104,000 GB300 GPUs at a 240MW AI campus in Texas, with services for Microsoft expected to start in Q3 2026 [4] - The site is planned to expand to 1.2GW over time, with Microsoft having an option for a second phase of approximately 700MW starting in late 2027 [4] - From Q1 2026, Nscale plans to deploy about 12,600 GB300 GPUs at the Start Campus in Sines, Portugal, aimed at EU customers [5] - In the UK, the Loughton AI Campus is expected to host around 23,000 GB300 GPUs from early 2027 in a scalable 50MW installation [5] - The Aker–Nscale joint venture in Narvik, Norway, is projected to deliver approximately 52,000 GB300 GPUs to Microsoft under a multi-year agreement [5] Strategic Importance - The partnership emphasizes the increasing demand for high-performance computing among enterprises to train and run AI models globally [6] - Nscale positions itself as a vertically integrated AI cloud provider, focusing on delivering sovereign-grade, sustainable capacity as next-generation workloads grow [7]