Analysis
SWI Compute, a new AI infrastructure provider launched by Euronext Amsterdam-listed SWI Group, has ordered roughly $1 billion of Nvidia hardware to build a European network of AI data centers, targeting 2.3 gigawatts of capacity by 2029, the company said November 2, 2026.
What SWI Compute does
SWI Compute builds and operates what it calls Nvidia AI factories -- large-scale data centers purpose-built for AI workloads -- and became an Nvidia Preferred Cloud Partner in August 2026. Its first facility, a 30-megawatt installation in the United Kingdom, will house nearly 7,000 Nvidia Rubin GPUs and 3,500 Vera CPUs and is due online by the end of 2027, meaning none of its planned capacity is live yet. The company's parent, SWI Group, is listed on Euronext Amsterdam, according to IntelligentCIO.
SWI Compute's capital commitments
SWI Compute's roughly $1 billion in ordered Nvidia systems is a purchase commitment, not a venture round -- the capital is coming from its listed parent rather than outside investors. SWI Group says it now holds 3.6 gigawatts of power capacity across 24 sites in Europe and the US, according to W.Media, the power base it's drawing on to support the new AI-factory buildout.
“The company's parent, SWI Group, is listed on Euronext Amsterdam, according to IntelligentCIO.”
Who's behind SWI Compute
SWI Compute has signed initial customer contracts, including one with what the company describes only as "a US-based AI inference provider," and has leased one facility to what a second outlet described as "a major hyperscaler" without naming it. Beyond its Nvidia partnership and parent company, no outside investors in SWI Compute itself have been named.
Europe's AI data center market and competitors
European data center investment totaled $59.84 billion in 2025 and is projected to reach $148.82 billion by 2031, according to market-research firm Arizton Advisory & Intelligence, as cited by W.Media -- the scale of buildout SWI Compute is now competing for. The most direct comparison:
- Nscale -- a UK-based AI cloud provider that has raised more than $1.7B in the past two months from Nvidia and other investors, signed a Microsoft contract worth up to $24B for roughly 200,000 Nvidia GB300 GPUs, and is building an initial Texas campus starting at 240 megawatts and scaling toward 1.2 gigawatts
Against that backdrop, SWI Compute's 2.3-gigawatt target by 2029 would be a substantial buildout but still a fraction of what Nscale alone has already committed to deliver. Pulse has tracked how AI megaround pricing keeps climbing even as credit markets grow more cautious, a dynamic that applies as much to infrastructure capex as to model-layer funding rounds.
What to watch next
Announced is not completed: SWI Compute's first facility isn't due online until the end of 2027, so the 2.3-gigawatt target is a multi-year plan, not current capacity, and the company has named only one customer and one unnamed hyperscaler lease so far. Power availability is the risk worth watching -- SWI Group's 3.6-gigawatt base across 24 sites, per W.Media, has to stretch across its other European commitments and SWI Compute's own AI-factory ambitions.
