SWI Compute exits stealth to build NVIDIA AI factories in Europe
SWI Compute has emerged from stealth to build and operate NVIDIA AI factories, starting in Europe. The new company of Euronext Amsterdam-listed SWI Group says it is backed by 3.6 gigawatts of power capacity that the group has spent six years securing across Europe and the US.
- SWI Compute emerged from stealth to build and operate NVIDIA AI factories, starting in Europe.
- It is backed by 3.6 GW of power capacity secured by SWI Group over six years across Europe and the US.
- The company became a Preferred NVIDIA Cloud Partner in August and ordered approximately $1 billion of NVIDIA systems.
- The first UK AI factory starts at 30 MW with close to 7,000 Rubin GPUs and is expected online by the end of 2027.
- The UK campus is planned to reach 530 MW, within a target of 2.3 GW of NVIDIA AI factory capacity by 2029.
A power-first approach to AI infrastructure
According to the company, SWI Compute became a Preferred NVIDIA Cloud Partner in August, ordered approximately $1 billion of NVIDIA systems and has finalized contracts with its first customers. Its first customer is a US-based AI inference platform that has contracted capacity.
The parent, SWI Group, began investing in land, power and shell in 2020. It now holds 3.6 GW of power capacity across Europe and the US, with five European campuses in Ireland, the UK, Denmark, Spain and Italy. One of them is already leased to a major hyperscaler. The group develops its own sites, from grid connection to commissioning.
Halil Mirakhmedov, CEO of SWI Compute, said: "Energy availability represents the primary bottleneck facing European AI infrastructure." He added that the market continues to surpass available capacity.
The UK site and the expansion plan
The first large-scale AI factory is in the UK. It starts with 30 MW, close to 7,000 Rubin GPUs and 3,500 Vera CPUs in NVIDIA Vera Rubin NVL72 racks, and is expected to come online by the end of 2027. The racks are connected by NVIDIA Quantum-X800 InfiniBand and NVIDIA Spectrum-X Ethernet networking.
These 30 MW are the initial phase of a UK campus planned to grow to 530 MW. The company calls it the first step towards 2.3 GW of NVIDIA AI factory capacity by 2029. Starting in Q4 2027, a next generation of AI campuses across Europe is scheduled to come online, with thousands of GPUs planned.
The first deployment runs NVIDIA Blackwell Ultra systems, with NVIDIA Vera CPU racks to follow. Delta Computers is the systems integrator, alongside storage partners, to deliver more than 8,000 GPUs.
Design based on NVIDIA reference architectures
Every SWI Compute site is built to NVIDIA reference architectures. Going forward, the company will design, build and operate its AI factories with the NVIDIA DSX AI Factory Platform, which codesigns compute, networking, power and cooling as one integrated system. Every new site from 2027, starting with the UK campus, will follow the NVIDIA Vera Rubin DSX AI Factory reference design. According to the release, on Vera Rubin NVL72 the DSX platform can enable up to 40% more GPU capacity within the same power budget.
SWI Compute argues that an AI factory earns on three things: tokens produced per megawatt, how many years it keeps producing them, and how many kinds of work it can take on. The same infrastructure, it says, can run every type of AI in every phase, from pretraining to agentic inference, as well as simulation, data processing and scientific computing.
Who it is aimed at
The company says capacity in Europe has been committed first to global hyperscalers and frontier labs, the buyers able to sign multi-year contracts at scale. AI natives with frontier ambitions and startup balance sheets have the hardest time securing it. SWI Compute says it is built for them, offering dedicated and on-demand access to NVIDIA AI factories for Europe's AI natives, AI labs and enterprises. The capacity sits close to users and data and under European jurisdiction.
Nico Caprez, Vice President, Global AI Infrastructure Growth at NVIDIA, said SWI Compute is building infrastructure that is "productive, durable and fungible". Max-Hervé George, Founder and CEO of SWI Group, described the group as a transatlantic, vertically integrated digital infrastructure company, "from the grid connection to the GPU".
For European AI developers, the plan points to more locally based GPU capacity, though the first UK systems are not expected online until the end of 2027.
