Nscale Signs $14B Microsoft Deal for 200K Nvidia GPUs

• UK-based Nscale to supply 200,000 Nvidia GPUs in a landmark $14 billion agreement.
• The chips will power AI data centers in the US, Portugal, Norway, and the UK starting in 2026.

On October 15, 2025 (UTC), Cryptopolitan reported that UK-based AI infrastructure firm Nscale secured a $14 billion agreement with Microsoft to deliver approximately 200,000 Nvidia GB300 GPUs. This deal ranks as one of the most significant AI infrastructure partnerships to date, highlighting the surging demand for advanced computing capabilities fueled by AI advancements.

The GPUs will be deployed across four hyperscale AI data centers located in Texas, USA; Portugal; Norway; and the United Kingdom. Of these, Texas will receive the largest allocation, with 104,000 GPUs allocated for a 240 MW campus. Deliveries to the Texas site are set to begin in the third quarter of 2026.

Portugal’s Start AI campus is slated to receive 12,600 GPUs, with deployment planned for the first quarter of 2026. In the United Kingdom, Nscale will construct a new AI data center at its Loughton Campus, which is expected to host 23,000 GPUs by early 2027. Meanwhile, Norway will receive 52,000 GPUs to support operations in Narvik, building on an existing joint venture between Nscale and Norwegian industrial company Aker ASA.

This agreement is poised to bolster Microsoft’s Azure AI workloads, which require vast computational capacity for training and inference of complex AI models. The four campuses are being designed for sustainability, integrating renewable energy sources and advanced modular cooling technologies to minimize environmental impact.

Nscale’s $14 billion deal comes on the heels of its recent $1.1 billion Series B funding round, led by Aker and backed by participants such as Nvidia, Dell, and Nokia. In a statement to the Financial Times, Nscale co-founder and CEO Josh Payne emphasized that the partnership establishes the company as a trusted provider for leading global tech firms. Payne also revealed ongoing plans for an initial public offering (IPO) by the end of 2026.

This collaboration signals a growing trend of large-scale investments in AI infrastructure, combining top-tier GPU technology with environmentally sustainable designs to meet the skyrocketing computational demands of the AI era.