Home 9 Computing 9 Nscale Plans 66,000 NVIDIA Rubin GPUs for SINES Data Campus

Nscale Plans 66,000 NVIDIA Rubin GPUs for SINES Data Campus

by | May 6, 2026

Microsoft partnership adds a second 200MW building in Portugal for AI infrastructure
Image: Nscale

LONDON, UK, May 6, 2026 – Nscale plans to expand its AI infrastructure program with Microsoft and Start Campus by adding more than 66,000 NVIDIA Rubin GPUs at the SINES Data Campus in Portugal, starting in late 2027. The expansion includes a second building and follows the deployment of more than 12,600 NVIDIA Blackwell Ultra GPUs for Microsoft at the same campus.

The agreement includes Nscale investments of €230 million in shared infrastructure and €465 million for the second 200MW building. It also extends Nscale’s Microsoft deployments in Norway, the U.K. and the U.S.

The companies cited rising AI infrastructure demand through 2030, with power availability and new data center capacity affecting deployment schedules. Start Campus is permitted for 1.2GW, giving the SINES site capacity for additional AI compute infrastructure.

“This partnership enables the deployment of next-generation AI compute at the scale and efficiency required for frontier workloads. Building on a proven foundation, the expanded deployment in Sines, Portugal creates one of the most advanced environments in Europe for high-density AI infrastructure. It also represents one of the largest AI infrastructure investments in Portugal’s history – and among the most significant in the EU – reflecting the surging demand we’re seeing for Nscale’s services,” said Josh Payne, CEO and founder of Nscale.

“Sines is one of Europe’s leading destinations for large-scale AI – strengthening Europe’s ability to support sovereign AI development with sustainability, resilience and long-term planning at its core. Building on our existing agreement, this expansion by Nscale will be one of Europe’s largest NVIDIA Vera Rubin NVL72 deployments. As we progress towards our 1.2GW vision, our focus remains on delivering secure, sustainable and globally connected infrastructure – underpinned by renewable energy and designed to support Europe’s long-term competitiveness in the AI era,” said Robert Dunn, CEO of Start Campus.

“This significant investment by Nscale accelerates the development of the SINES Data Campus. Underpinned by Davidson Kempner’s continued commitment, the Start Campus platform is well positioned to scale – reinforcing Portugal’s emergence as an AI leader in Europe,” said Daniel Boehm, partner at Davidson Kempner, shareholder of Start Campus.

The expansion adds capacity for NVIDIA Vera Rubin NVL72 systems used in frontier AI workloads. The project adds AI compute capacity in Europe and expands Portugal’s role in high-density data center infrastructure.

Source: Nscale

About Start Campus

Start Campus is a data center developer and operator based in Lisbon, Portugal. Founded in 2020, the company is developing the SINES Data Campus in Sines, Portugal, with planned capacity of 1.2 gigawatts. It provides powered shell, turnkey and data center services for AI, cloud, high-performance computing, colocation, hyperscale and enterprise infrastructure customers. Its Sines facilities use renewable power and seawater cooling to support dense computing workloads. The first facility, SIN01, has 33 megawatts of live capacity. The planned SIN02 facility is listed at 200 megawatts. Start Campus works with connectivity, energy and data center technology partners.

About Nscale

Nscale Ltd., founded in 2024, is a UK-based technology company specializing in AI-optimized cloud infrastructure and high-performance computing (HPC) solutions. Headquartered in London, with data centers in Glomfjord, Norway, Nscale offers a vertically integrated platform encompassing public and private cloud services, serverless inference, GPU clusters, and AI development tools. Its infrastructure is powered by renewable energy, emphasizing sustainability. Nscale serves various industries, including education, finance, government, healthcare, legal, manufacturing, software, and telecommunications, providing scalable solutions for AI model training, inference, and development. In January 2025, the company secured $155 million in Series A funding led by Sandton Capital Partners, supporting its expansion across Europe and North America.