15.10.2025 16:25

In a landmark move for the artificial intelligence sector, UK-based firm Nscale has forged a pivotal, multi-billion-dollar partnership with Microsoft to deploy up to 200,000 of Nvidia’s state-of-the-art GB300 GPUs. This monumental agreement, valued at an estimated $14 billion and facilitated in collaboration with Dell Technologies and a joint venture involving Aker ASA, will see the advanced chips integrated into Microsoft's data centers across the United States and Europe.

Central to this initiative is the allocation of 104,000 Nvidia GB300 units to a new 240 MW hyperscale Microsoft AI data center in Texas, a detail confirmed by Nscale in a public statement. With initial shipments slated for the third quarter of 2026, the campus, leased from Ionic Digital, has ambitious plans for future growth, with its capacity projected to eventually scale to an immense 1.2 GW.

The transatlantic expansion continues with a significant deployment in Europe, commencing with 12,600 GPUs destined for the Start AI campus in Portugal in the first quarter of 2026 to bolster Microsoft's Azure AI services for regional customers. Furthermore, Microsoft is set to establish a new flagship AI data center at Nscale's Loughton Campus in the United Kingdom, which will become the nation's largest such facility upon its projected completion in early 2027, housing approximately 23,000 GB300 GPUs and featuring an initial 50 MW capacity expandable to 90 MW.

According to information from internet sources, this massive infrastructural investment is a direct response to the soaring and unprecedented demand for high-capacity AI training and inference capabilities throughout the region.