Startup company Nscale, which specializes in AI cloud computing infrastructure, has just signed a large-scale agreement with Microsoft to deploy 200,000 Nvidia GB300 GPUs at data centers in Europe and the United States.
This is considered one of the largest AI infrastructure deals in 2025, marking the rapid expansion of Nscale after only one year of establishment.
According to information from Nscale, three data centers in Europe and one in Texas (USA) will be equipped with these huge GPUs.
About 104,000 GPUs will be transferred to the Ionic Digital-operated facility in Texas over the next 12-18 months, with plans to increase the total capacity here to 1.2 gigawatt.
Along with that, 12,600 GPUs will be deployed at the Start campus data center in Sines, Portugal in the first quarter of 2026.
In addition, Nscale continues to expand to the UK and Norway, as it is expected to send 23,000 GPUs to the Loughton (UK) facility in 2027 and 52,000 GPUs to Microsoft's AI center in Narvik (Na Uy).
These projects are implemented through a joint venture between Nscale and Aker, a strategic investor and long-term partner of the company.
This deal affirms Nscales position as a partner chosen by the worlds leading technology leaders, said Josh Payne, founder and CEO of Nscale.
Mr. Payne also emphasized that the deployment of GPU infrastructure at this scale is a testament to the company's capacity, sustainability and global vision.
Notably, Nscale was only founded in 2024, but has raised more than $1.7 billion from strategic partners such as Aker, Nokia and Nvidia, along with large investment funds such as Point72, G Squared and Sandton Capital Partners.
Payne revealed that the company is considering IPO by the end of 2026 at the earliest.
The event comes as the global GPU market heats up. Last week, OpenAI announced the purchase of 6 gigawatt chips from AMD, and reached a deal that would allow Nvidia to invest $100 billion in exchange for 10 gigawatt GPUs.
With the deal with Microsoft, Nscale has joined a group of big players shaping the global AI infrastructure race, where GPU deployment capabilities become a strategic weapon for the modern technology industry.
