AI hyperscaler startup Nscale has finalized a major infrastructure agreement with Microsoft to deploy approximately 200,000 Nvidia GB300 GPUs across four global data centers. This strategic expansion, announced Wednesday, targets facilities across the United States and Europe to bolster the next generation of AI compute capacity.
Global Infrastructure Expansion
The GPU rollout will be executed through Nscale’s proprietary operations and a joint venture with the investment firm Aker. The deployment schedule is set to unfold over the coming months:
- Texas, U.S.: 104,000 GPUs will be installed in a facility leased by Ionic Digital over the next 12 to 18 months, with plans to scale the site to 1.2 gigawatts.
- Sines, Portugal: 12,600 GPUs are slated for the Start Campus data center, beginning in Q1 2026.
- Loughton, England: 23,000 GPUs will be deployed to the company’s campus starting in 2027.
- Narvik, Norway: The remaining 52,000 GPUs will be integrated into Microsoft’s AI campus.
Scaling AI Capabilities
Josh Payne, founder and CEO of Nscale, emphasized that this deal positions the firm as a primary partner for global technology leaders. “Few companies are equipped to deliver GPU deployments at this scale,” Payne stated in a company press release. “We have the experience and the global pipeline to set a new standard for AI infrastructure delivery.”
Rapid Market Ascent
Despite being founded only in 2024, Nscale has rapidly solidified its market position. The company has secured over $1.7 billion in funding from high-profile backers, including Nokia, Nvidia, Sandton Capital Partners, G Squared, and Point72. Following this momentum, Payne indicated to the Financial Times that Nscale is currently evaluating an IPO as early as the end of 2026.
The Competitive Landscape
The agreement reflects an intensifying race for AI hardware, with major players aggressively securing capacity. Recent market shifts include OpenAI’s move to acquire six gigawatts of chips from AMD, alongside a substantial investment deal with Nvidia involving 10 gigawatts of chip capacity. Nscale’s latest commitment to Microsoft underscores the critical demand for massive-scale GPU infrastructure as the industry prepares for the next phase of AI development.