Civo has launched the first of 40 planned edge data centers in the UK, designed to provide high-performance capacity for sovereign AI workloads. The facilities will utilize direct-to-chip liquid cooling and NVIDIA Vera Rubin systems to support high-density inference processing.
The deployment of localized, high-density compute reduces reliance on centralized hyperscalers and international supply chains for sensitive AI workloads. By placing NVIDIA Vera Rubin systems physically closer to UK organizations, Civo addresses latency and data sovereignty requirements essential for real-time inference applications. This move signals a broader shift in the streaming and AI ecosystem toward distributed edge architectures that prioritize local compliance over globalized cloud models. Watch for the completion of the remaining 34 sites over the next three years to see if Civo can maintain its performance lead as competitors upgrade their own UK-based silicon footprints.
Civo's distributed edge strategy is grounded in measurable demand from UK enterprises. At the Civo Navigate event in London, CEO Mark Boost revealed that the company's annual survey of 1,000 UK IT leaders found 73 percent now rate AI sovereignty as a strategic priority, up from 61 percent the previous year. That data point underpins the business case for 40 geographically dispersed sites rather than a single hyperscale campus, and it positions Civo against US-controlled cloud providers whose UK data centres remain subject to the US CLOUD Act. Boost explicitly framed sovereignty as a question of ownership and legal jurisdiction, not merely physical location of hardware.
The regulatory and geopolitical framing is central to Civo's positioning. The company argues that data stored in a UK building owned by a foreign entity is still exposed to overseas legal compulsion, a concern that has intensified as governments tighten data-residency requirements for public-sector and critical-infrastructure AI workloads. Civo's pitch to international organisations is a compliant UK home for services and data, with the Hertfordshire site opening at eight megawatts and expanding to 38 megawatts on the same plot, while five further locations in Lancashire, Greater Manchester, London, and Oxfordshire are secured to reach 150 megawatts within three years. The remaining sites are planned over the following three years to reach the one-gigawatt target.
On the technical side, Civo is among the first UK cloud providers to offer the NVIDIA Vera Rubin platform, which succeeds Blackwell with up to five times the inference performance. Each NVL72 rack pairs 72 Rubin GPUs with 36 Vera CPUs, delivering 288 GB of HBM4 memory per GPU and up to 3,600 PFLOPS of NVFP4 inference throughput. Civo is already taking orders for Vera Rubin at early-access pricing starting at $10 per hour, alongside existing capacity on NVIDIA B300, B200, H100, and H200 instances. The entire build is 100 percent direct-to-chip liquid cooled, a necessity as GPU density rises with each successive generation. Oracle sovereign AI framework provides a similar approach to data residency for enterprises.
Civo has launched the first of 40 planned UK edge data centres, aiming to provide one gigawatt of capacity for sovereign AI. By utilizing NVIDIA Vera Rubin systems and direct-to-chip liquid cooling, Civo offers localized, high-density compute that helps UK organizations address data sovereignty requirements and reduce reliance on centralized hyperscalers.
Civo is planning to build 40 edge data centres across the UK over the next three years to reach a one-gigawatt capacity target.
The new facilities feature NVIDIA Vera Rubin NVL72 systems, which include 72 Rubin GPUs and 36 Vera CPUs, providing 288GB of HBM4 memory per GPU.
Civo is addressing concerns regarding data residency and legal jurisdiction, as 73 percent of UK IT leaders now rate AI sovereignty as a strategic priority to avoid reliance on foreign-owned cloud providers.
The entire build utilizes direct-to-chip liquid cooling, which is necessary to support the high-density inference performance of the new GPU generations.
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source