AI hyperscaler Nscale has entered into a significant partnership with Microsoft to supply Nvidia AI hardware across several data centers.
On Wednesday, the AI cloud firm revealed it had reached an agreement with Microsoft to provide around 200,000 Nvidia GB300 GPUs to three European data centers and one in the United States.
These GPUs will be distributed through Nscale’s own facilities as well as through a partnership with Aker, an investment firm backing Nscale.
Of the total GPUs, 104,000 are slated for a Texas data center operated by Ionic Digital over the next year to year and a half. Nscale aims to expand its capacity at this site to 1.2 gigawatts, according to the company.
Additionally, Nscale will allocate 12,600 GPUs to the Start Campus facility in Sines, Portugal, beginning in the first quarter of 2026.
This arrangement builds upon earlier collaborations with both Microsoft and Aker for data centers in Norway and the UK. Nscale intends to deliver 23,000 GPUs to its Loughton, England location starting in 2027, while the remaining 52,000 GPUs will be sent to Microsoft’s AI facility in Narvik, Norway.
“This partnership solidifies Nscale’s reputation as a preferred collaborator for the world’s leading technology companies,” said Josh Payne, Nscale’s founder and CEO, in a press statement. “Very few organizations can manage GPU rollouts of this magnitude, but our expertise and global network make it possible.”
Payne’s assertion is ambitious, especially since Nscale was only established in 2024. Since its inception, the company has secured over $1.7 billion in funding from strategic partners such as Aker, Nokia, and Nvidia, as well as investors like Sandton Capital Partners, G Squared, and Point72. Payne mentioned to the Financial Times that Nscale is considering a public offering as soon as late next year.
“The speed at which we’ve scaled up our operations highlights our preparedness, our focus on efficiency and sustainability, and our dedication to offering clients the most advanced solutions. This is a strong indication that Nscale is redefining the standards for the next generation of AI infrastructure,” Payne added in the release.
Recently, GPU procurement deals have accelerated. Last week, OpenAI disclosed it would acquire chips from AMD totaling six gigawatts. OpenAI also recently finalized an agreement with Nvidia, where Nvidia will invest up to $100 billion in exchange for 10 gigawatts of chips this September.
