SWI Compute emerged from stealth on Monday to build and operate NVIDIA AI factories, launching first in Europe on a power base its parent has spent six years assembling. The company, part of Euronext Amsterdam-listed SWI Group, says it is backed by 3.6 gigawatts of power capacity secured across Europe and the United States, became a Preferred NVIDIA Cloud Partner in August, has ordered approximately 1 billion dollars of NVIDIA systems, and has finalized contracts with its first customers.
The first large-scale site, in the United Kingdom, starts at 30 megawatts with close to 7,000 Rubin GPUs and 3,500 Vera CPUs in NVIDIA Vera Rubin NVL72 racks, connected by NVIDIA Quantum-X800 InfiniBand and NVIDIA Spectrum-X Ethernet networking. It is expected to come online by the end of 2027. Those 30 megawatts are only the initial phase of a UK campus planned to grow to 530 megawatts, and the company describes it as the first step toward 2.3 gigawatts of NVIDIA AI factory capacity by 2029.
Power, not chips, is the pitch. SWI Group began investing in land, power and shell in 2020 and now holds capacity across 24 sites, with five European campuses in Ireland, the United Kingdom, Denmark, Spain and Italy, one of them already leased to a major hyperscaler. The group develops its own sites from grid connection to commissioning, so SWI Compute starts with contracted power and systems already on order rather than a plan to find both later.
Every site is built to NVIDIA reference architectures. From 2027, new sites, starting with the UK campus, will follow the NVIDIA Vera Rubin DSX AI Factory reference design, using the full-stack NVIDIA DSX platform, which codesigns compute, networking, power and cooling as one system. The company says that on Vera Rubin NVL72, DSX can enable up to 40 percent more GPU capacity within the same power budget. The first deployment itself runs NVIDIA Blackwell Ultra systems, with Vera CPU racks to follow, and systems partners selected from NVIDIA's OEM and ODM ecosystem, including Delta Computers as systems integrator, are set to deliver more than 8,000 GPUs.
SWI Compute is aiming squarely at buyers it says have been squeezed out: Europe's AI natives, labs and enterprises that need dedicated or on-demand capacity close to their users and data, under European jurisdiction, while existing capacity has gone first to global hyperscalers and frontier labs able to sign multi-year contracts at scale. Its first customer, a U.S.-based AI inference platform, has already contracted capacity.
"Energy availability represents the primary bottleneck facing European AI infrastructure," said Halil Mirakhmedov, chief executive of SWI Compute, noting that just two months after joining the NVIDIA Cloud Partner program the team had secured the hardware orders and initial client contracts. Nico Caprez, NVIDIA's vice president for global AI infrastructure growth, called AI factories essential infrastructure that turns regional energy into intelligence and economic opportunity.
For enterprise IT buyers, the launch is another sign that AI capacity is being sold less like cloud rental and more like industrial capacity: secured power, reference designs and networking fabrics decided years ahead. Starting in the fourth quarter of 2027, SWI Compute's next generation of European campuses is scheduled to come online with thousands more GPUs planned, if the power-first bet pays off.