Under the deal, Nscale will deliver around 104,000 NVIDIA GPUs to a 240-megawatt data center campus in Texas, beginning in the third quarter of 2026. The facility, leased from Ionic Digital, is engineered to scale up to 1.2 gigawatts of total capacity. Microsoft also retains an option for a second development phase that would add 700 megawatts by late 2027, significantly enhancing its U.S. cloud and AI capabilities.
In Europe, Nscale’s rollout includes 12,600 GPUs at the Start Campus in Sines, Portugal, slated to begin operations in early 2026. Another 52,000 GPUs will be deployed at a new data center in Narvik, Norway, through Nscale’s joint venture with Aker ASA. Additionally, Microsoft’s Azure cloud division is set to integrate roughly 23,000 GPUs at Nscale’s Loughton AI Campus in the United Kingdom by 2027.
Nscale CEO Josh Payne described the multiregional agreement as a milestone that reaffirms the company’s status as a premier partner for the world’s top cloud and AI developers. “This collaboration positions Nscale at the heart of global AI infrastructure growth, ensuring high-efficiency and sustainable computing capacity for next-generation innovation,” Payne said.
Microsoft executive Jon Tinter emphasized that the partnership aligns with the tech giant’s broader goals of scalability, sustainability, and international AI development. “Our collaboration with Nscale expands Microsoft’s global AI footprint while advancing our commitment to responsible, energy-efficient infrastructure,” Tinter stated.
The expansive initiative further strengthens Microsoft’s position in the rapidly intensifying AI arms race, as demand for advanced computing power surges across industries from cloud services to generative AI research.