Enterprise AI
Nebius Commits £1.7 Billion to UK NVIDIA AI Infrastructure Expansion
The AI cloud company is adding three new NVIDIA-powered sites to reach 65 MW capacity by 2027, supporting enterprise workloads and aligning with the UK government's sovereign compute strategy.
Nebius is an AI cloud company that has announced plans to invest approximately £1.7 billion in expanding its NVIDIA-powered infrastructure across the United Kingdom.
Nebius is an AI cloud company that has announced plans to invest approximately £1.7 billion in expanding its NVIDIA-powered infrastructure across the United Kingdom. The commitment covers three additional sites that will use the latest generations of NVIDIA’s full-stack AI factory platform technology. These sites join an existing deployment launched in November 2025 that already runs NVIDIA Blackwell Ultra infrastructure. The overall build-out brings the company’s total UK capacity to 65 MW once fully operational in 2027. The investment reflects sustained demand for localized high-performance compute from both startups and established enterprises operating in regulated sectors.
What is the background to Nebius’s UK operations?
Nebius has maintained a presence in the United Kingdom for several years and has already invested more than £200 million in local facilities. One of those facilities is a data centre located in Surrey that was fitted out to support early AI workloads. The company’s initial UK site began operations with NVIDIA Blackwell Ultra hardware in November 2025. That launch provided the technical foundation for subsequent expansions. The current announcement scales those capabilities significantly while maintaining the same technical standards and operational practices.
The United Kingdom has positioned itself as a destination for AI development through a combination of policy support and existing technology clusters. Nebius’s decision to increase capacity responds directly to that environment. Enterprises in the country require reliable access to training and inference resources without transferring data outside national borders. The expansion therefore addresses both commercial demand and strategic priorities around data residency.
What exactly does the new £1.7 billion investment include?
The £1.7 billion commitment funds three new deployments of NVIDIA infrastructure at separate UK locations. Each deployment will incorporate the most recent generations of NVIDIA’s end-to-end AI factory platform. When these sites reach full operation alongside the existing facility, total power capacity across all four sites will equal 65 MW. The timeline for completion targets 2027. The investment also includes additional cloud capabilities intended for agentic and enterprise AI use cases.
The three new sites will be constructed to the same technical specifications that govern the first UK deployment. This uniformity allows customers to move workloads between facilities without re-engineering applications. The full-stack approach integrates NVIDIA hardware, software, and networking components into a single managed platform. Such integration reduces the complexity that enterprises typically encounter when assembling AI infrastructure from multiple vendors.
What technical specifications define the new sites?
The infrastructure relies on NVIDIA’s full-stack AI factory platform technology. This includes the latest GPU generations optimized for both training and inference workloads. The first UK site already demonstrated Blackwell Ultra performance, and the new sites will extend that capability. Power delivery and cooling systems are sized to support sustained operation at the combined 65 MW level. The architecture supports the high interconnect bandwidth required for large-scale distributed training jobs.
| Aspect | Details | Timeline |
|---|---|---|
| Total Investment | Approximately £1.7 billion | Ongoing to 2027 |
| New Sites | Three deployments | 2026-2027 |
| Combined Capacity | 65 MW | 2027 |
| Existing UK Investment | More than £200 million | Prior to 2026 |
| First Deployment | NVIDIA Blackwell Ultra | November 2025 |
Which enterprises are already using the infrastructure?
Revolut relies on Nebius to run training and inference workloads for its financial crime detection systems and customer support agents. The platform supplies the scale required to serve 75 million customers while meeting the operational standards expected of a regulated financial institution. Prima Mente uses the same infrastructure to train biological foundation models. Both organizations cite the combination of performance, economics, and proximity to their teams as decisive factors.
The customer base illustrates the range of workloads that the expanded capacity will support. Financial services require low-latency inference at high volume. Biological research demands large-scale training runs that benefit from dense GPU clusters. The infrastructure design accommodates both patterns within the same facilities. This versatility increases utilization rates and improves the economics for the cloud operator.
- First UK deployment launched with NVIDIA Blackwell Ultra in November 2025
- Announcement of £1.7 billion commitment for three additional sites
- Ramp-up of all four sites to 65 MW total capacity by 2027
- Onboarding of enterprise customers including Revolut and Prima Mente
- Alignment with UK Government AI Opportunities Action Plan
What are the market and stakeholder implications?
The expansion increases the supply of sovereign AI compute available inside the United Kingdom. Companies that previously faced capacity constraints or had to use overseas resources can now access equivalent performance domestically. This change reduces latency for inference applications and simplifies compliance with data protection rules. The presence of additional NVIDIA-based capacity also strengthens the local ecosystem of AI developers and service providers.
Stakeholders in the UK technology sector view the investment as validation of the country’s AI strategy. The government’s AI Opportunities Action Plan explicitly calls for greater domestic compute resources. Nebius’s commitment supplies concrete capacity that matches those policy goals. At the same time, the move positions the UK more competitively against other European nations that are pursuing similar infrastructure programs.
How have officials and partners reacted to the announcement?
UK AI Minister Kanishka Narayan stated that the investment supplies the compute needed to train, test, and run advanced systems inside the country. NVIDIA Vice President EMEA Paolo Guglielmini noted that the sites give British companies access to the best performance and economics available from the NVIDIA platform. These statements emphasize both the technical quality of the infrastructure and its alignment with national priorities.
The UK is one of the places where AI is being built, deployed, and adopted at the same time — by startups, by enterprises, and by the public sector. The work is happening here and the demand is here. And we are also here for the long run.Arkady Volozh, founder and CEO of Nebius
What comes next for Nebius and the UK AI sector?
Nebius plans to continue scaling its UK footprint beyond the current 65 MW target as demand grows. The company has indicated it will more than triple its UK headcount to support the expanded operations. Additional cloud features for agentic AI and enterprise workloads are also under development. These steps will further integrate the UK sites into Nebius’s global network while preserving local data residency.
The broader UK AI sector is expected to benefit from the increased availability of domestic compute. Startups and research institutions that previously competed for limited resources will gain additional options. Regulated industries such as finance and healthcare can accelerate their AI adoption without compromising on security or compliance. The combination of policy support and private investment creates a reinforcing cycle that attracts further capital and talent to the country.
Nebius’s long-term commitment signals confidence in the durability of UK demand. The company’s existing customer relationships with Revolut and Prima Mente demonstrate practical use cases that justify the capital expenditure. As more organizations adopt similar AI strategies, the 65 MW capacity is likely to be fully utilized well before 2027. This utilization rate will in turn support additional investments in networking, storage, and software layers that complete the AI factory model.
The overall trajectory points toward a more self-sufficient European AI infrastructure landscape. The UK expansion contributes directly to that objective by anchoring significant NVIDIA capacity on European soil. Other nations observing the development may accelerate their own programs, leading to a wider distribution of compute resources across the continent. For enterprises, the result is greater choice and resilience when planning AI deployments that must meet both performance and regulatory requirements.