UK’s Nscale Signs $14 Billion Deal with Microsoft to Deliver 200,000 Nvidia AI Superchips Across US and Europe

London, UK – October 15, 2025 – In a monumental agreement that underscores the global race to build the infrastructure of the artificial intelligence era, UK-based AI hyperscaler Nscale has signed an expanded multi-billion-dollar deal with Microsoft to deploy approximately 200,000 NVIDIA GB300 superchips across data centers in the United States and Europe. The deal, estimated to be worth up to $14 billion in revenue for the British firm, is one of the largest AI infrastructure contracts ever signed and a massive victory for the relatively young company.

The agreement sees Nscale—which specializes in building and operating high-performance compute capacity for AI—solidifying its role as a critical partner for Microsoft’s massive global AI buildout, which powers its Azure AI services and Copilot products. The move provides Microsoft with a crucial injection of the latest-generation, highly sought-after AI hardware from Nvidia, ensuring it can keep pace with the exponential demand for AI training and inference capacity.

The Geo-Strategic AI Deployment

The sheer scale of the deployment spans four countries across two continents, cementing the Anglo-American partnership in the field of cutting-edge technology. The total count of nearly 200,000 NVIDIA GB300 GPUs will be delivered through Nscale’s owned and operated facilities, as well as its joint venture with Norway’s Aker ASA.

The key deployments include:

• Texas, US (Hyperscale AI Campus): The largest portion of the chips, approximately 104,000 NVIDIA GB300 GPUs, are destined for a massive $\sim$240MW hyperscale AI campus in Texas. The phased delivery is expected to begin in the third quarter of 2026. This facility, leased from Ionic Digital, is part of Nscale’s long-term plan to scale the site to an astonishing 1.2 GW, with Microsoft holding an option on a second 700MW phase beginning in late 2027.

• Narvik, Norway (Aker-Nscale Joint Venture): Building on an earlier agreement, the joint venture will deliver approximately 52,000 NVIDIA GB300 GPUs from its hyperscale AI campus in the Arctic Circle. This location leverages access to abundant, low-cost renewable hydroelectric energy, aligning with Microsoft’s commitment to sustainable scaling.

• Sines, Portugal (Start Campus): In Europe, Nscale will deploy approximately 12,600 NVIDIA GB300 GPUs at the Start Campus data center, with deployment commencing in Q1 2026. This strategically located capacity will support Microsoft’s Azure AI workloads and help provide sovereign AI solutions for its European customers within the EU.

• Loughton, UK (AI Campus): The deal also bolsters Nscale and Microsoft’s earlier plans to deliver the UK’s largest NVIDIA AI supercomputer. This 50MW facility, scalable to 90MW, will house around 23,000 NVIDIA GB300 GPUs and is expected to come online in Q1 2027, playing a vital role in providing Microsoft’s Azure services across the United Kingdom.

This expansive and distributed deployment strategy highlights the critical need for global infrastructure that is not only powerful but also geographically diverse, ensuring resilience and compliance with various data sovereignty regulations across different regions.

The Rise of Nscale

For Nscale, the deal serves as a dramatic validation of its unique, high-capital business model and catapults the London-based company into the front ranks of global AI infrastructure providers. Founded in 2024 as a spinout from a crypto mining company, Nscale has quickly pivoted to become a vertically integrated AI cloud provider, offering GPU clusters, data centers, orchestration, and AI services.

The company has successfully attracted significant investment and strategic backing from industrial and tech giants. In the past month alone, Nscale secured $1.5 billion in new funding, raising its valuation to an estimated $3.1 billion. Investors include Norwegian investment firm Aker ASA, Nvidia (which recently invested £500 million and called Nscale a “national champion for the UK”), Dell Technologies, and Nokia, among others. This funding surge and the massive Microsoft contract have accelerated Nscale’s plans for a potential Initial Public Offering (IPO) in late 2026.

“This agreement confirms Nscale’s place as a partner of choice for the world’s most important technology leaders,” said Josh Payne, founder and CEO of Nscale. “Few companies are equipped to deliver GPU deployments at this scale, but we have built the global pipeline to do so. It’s a clear signal that Nscale is setting a new standard for how the next wave of AI infrastructure will be delivered.”

Microsoft’s AI Infrastructure Push

For Microsoft, this partnership is a crucial component of its broader strategy to rapidly scale its cloud compute capacity to support the explosion of generative AI services like Azure OpenAI Service and Microsoft Copilot. By partnering with specialist providers like Nscale, Microsoft can outsource the complex, capital-intensive, and power-hungry process of building hyperscale AI data centers, while focusing on integrating the capacity into its cloud platform.

Jon Tinter, President of Business Development and Ventures at Microsoft, commented on the collaboration: “Together with Nscale, Microsoft is delivering cutting-edge AI infrastructure for our customers. Our agreement, announced today, demonstrates our commitment to ensuring that our products are available globally with sustainability and scalability in mind. Nscale is an ideal partner for this mission, given its deep expertise in providing AI infrastructure services at sca²le.”

A significant aspect of the deployment is the focus on sustainability. Nscale noted that all forthcoming campuses have been designed to incorporate modular cooling systems and utilize renewable energy sources—a non-negotiable factor as the energy demands of large AI clusters continue to soar.

The Broader Impact on the AI Race

The announcement is a powerful indicator of several key trends shaping the global technology landscape:

1. Outsourcing Compute: Major hyperscalers are increasingly relying on specialized AI infrastructure partners to meet the insatiable demand for GPUs amid supply constraints.

2. The New Sovereign Compute: The deployment across the UK, Norway, and Portugal highlights the push for sovereign AI capacity—AI infrastructure located within national or regional borders to address data sovereignty and security concerns for government and enterprise customers.

3. The Power Challenge: With demand for AI compute capacity nearly selling out the pipeline for the next year, the key challenge in the AI race has shifted from simply acquiring chips to securing the vast amounts of power and cooling required to run them. Nscale’s strategic choice of locations, particularly in the Nordics, reflects this reality.

The $14 billion deal not only secures massive compute power for one of the world’s leading technology companies but also marks the coming-of-age for Nscale, turning a London-based start-up into a central, global player in the foundation of the AI revolution.