AI infrastructure startup Nscale has entered a major partnership with Microsoft to supply a significant number of Nvidia GPUs across multiple global data centers. This move marks a notable expansion in Nscale’s ambitions to become a leading provider of AI compute infrastructure.
The deal involves deploying around 200,000 Nvidia GB300 GPUs to four key data center locations—three in Europe and one in the United States. These GPU units will be distributed via Nscale’s own operations and a joint venture with Aker, an investment firm and existing backer of the company.
Of the total, approximately 104,000 GPUs are designated for a Texas data center, which is currently leased by Ionic Digital. These will be installed over a period of 12 to 18 months, and Nscale has indicated it aims to expand its infrastructure at this site to a 1.2-gigawatt capacity.
In Europe, one of the major deployment sites will be the Start Campus data center in Sines, Portugal, where 12,600 GPUs are expected to be installed starting in the first quarter of 2026. This site forms a key part of Nscale’s European presence.
Nscale is also moving forward with earlier plans to support data centers in both Norway and the United Kingdom. In Loughton, England, the company will begin delivering 23,000 GPUs starting in 2027, further strengthening its footprint in the UK.
Meanwhile, the remaining 52,000 GPUs from the agreement will be directed to Microsoft’s AI facility in Narvik, Norway. These European deployments underscore the scale and geographic reach of Nscale’s growing infrastructure network.
Nscale’s CEO and founder, Josh Payne, stated that the partnership validates the company’s emergence as a trusted provider for leading AI and cloud firms. He emphasized that few organizations have the capability to deliver GPU clusters at this magnitude.
Despite being founded only in 2024, Nscale has made rapid progress. The company has already secured more than $1.7 billion in funding from strategic investors such as Aker, Nokia, and Nvidia, along with financial backers like Sandton Capital Partners, G Squared, and Point72.
Payne also mentioned that the company is exploring a potential initial public offering (IPO) as early as late next year, according to statements made in a recent interview. This signals strong confidence in Nscale’s growth trajectory and investor interest.
As the demand for AI infrastructure accelerates, major GPU supply agreements are becoming more common. Just last week, OpenAI announced a large-scale chip purchase from AMD totaling six gigawatts, and in September, it finalized a deal with Nvidia that could be worth up to $100 billion in exchange for access to 10 gigawatts of AI chips.
Also Read:
Walmart Partners with OpenAI to Bring ChatGPT into Shopping
Mozilla’s Firefox adds Perplexity AI answer as a New Search Engine Option
