Nscale Commits Over €695 Million to Expand NVIDIA-Powered AI Infrastructure for Microsoft in Portugal

New updates have been reported about Nscale.

Nscale is significantly scaling its European AI infrastructure footprint, committing to deploy more than 66,000 NVIDIA Rubin GPUs for Microsoft at Start Campus’s SINES Data Campus in Portugal, building on an initial rollout of over 12,600 NVIDIA Blackwell Ultra GPUs at the site’s first building. The new Rubin-based systems, including one of the EU’s largest NVIDIA Vera Rubin NVL72 deployments, are intended to support frontier AI workloads for Microsoft’s European customer base and materially increase Nscale’s capacity to serve hyperscale demand.

To enable this expansion, Nscale plans over €695 million of additional investment, including €230 million in shared infrastructure and €465 million in a second 200MW facility at SINES, positioning the project among the largest AI infrastructure builds in both Portugal and the EU. Management frames the deal as a response to rapidly rising AI compute demand constrained by power and capacity, with SINES offering a fully permitted runway to 1.2GW, underpinned by renewable energy and high-density, liquid-cooling-ready design.

The enlarged deployment deepens Nscale’s strategic relationship with Microsoft, complementing existing projects in Norway, the U.K., and the U.S., and reinforces its vertically integrated model spanning energy, data centers, GPU compute, and software. CEO Josh Payne describes the Portugal build-out as creating one of Europe’s most advanced environments for large-scale AI training and inference, while Start Campus and its backers highlight the project’s role in establishing Portugal as a secure, sustainable gateway for Europe’s AI economy.

For executives and investors, the initiative signals Nscale’s intent to be a core infrastructure provider for sovereign and enterprise AI in Europe, locking in long-duration demand from a top-tier cloud customer while anchoring growth on a scalable, fully permitted campus. The project also illustrates how access to power, connectivity, and advanced cooling is becoming a key differentiator in AI infrastructure, and it positions Nscale to benefit from expected tightness in high-end GPU capacity through 2030.