Nebius Announces £1.7B UK Nvidia Infrastructure Expansion

Nebius announced it is investing approximately £1.7B to deploy three new Nvidia infrastructure sites in the U.K., according to TipRanks and Seeking Alpha. TipRanks reports the three new sites will deploy the latest generations of Nvidia's full-stack AI platform and that combined capacity will reach 65 MW when fully ramped up in 2027. TipRanks further reports Nebius launched its first U.K. deployment of Nvidia Blackwell Ultra in November 2025 and is actively hiring engineering and R&D talent in the U.K., complementing its Amsterdam headquarters. Seeking Alpha reports Nebius shares rose about 4% premarket and Nvidia shares rose about 2% on the announcement.

Companies expanding regional AI capacity typically cite two technical drivers: access to lower-latency, high-throughput GPU clusters and control over power and cooling at scale. For practitioners, an additional 65 MW of locally sited Nvidia hardware materially increases regional capacity for large-model training and inference workloads, reducing queuing pressure on hyperscale public clouds. Deployments described as using Nvidia's "full-stack" platform imply combination of GPU compute, networking (such as NVLink/NIC fabrics), and software stack components, which together influence throughput, model parallelism choices, and cost-per-token for inference and training.

Industry reporting frames the Nebius build-out as increasing domestic compute capacity that aligns with the U.K. Government's AI Opportunities Action Plan, per TipRanks. Observed patterns in similar expansions show local compute availability can accelerate adoption by enterprises and research institutions that have data residency, latency, or regulatory constraints. For the broader ecosystem, more regional capacity creates more options for managed private clusters, partner-colocated services, and customers seeking non-U.S. cloud compute footprints.

• •Track advertised service offerings and pricing models from Nebius as sites come online, since commercial differentiation will hinge on SLAs, networking options, and availability of managed MLOps tooling.
• •Monitor reported site power ramp timelines versus the 2027 target reported by TipRanks, because power availability and grid integration are common bottlenecks for large GPU deployments.
• •Observe local hiring and partnerships announced in the U.K.; TipRanks reports active recruitment of engineering and R&D staff, which affects ecosystem capability for managed services and enterprise integrations.

This is a notable regional infrastructure expansion that increases U.K. onshore GPU capacity and may relieve some demand pressure for regional customers and R&D teams. Industry observers and practitioners will watch service-level details, power ramp timing, and how Nebius packages Nvidia hardware and software into accessible offerings.