Schneider tests NVIDIA DSX twin at SINES

Schneider Electric, Nscale and Start Campus have brought the NVIDIA Omniverse DSX Blueprint into a live AI factory environment through a digital twin pilot at the SINES Data Campus in Portugal. The Immersive Twin connects infrastructure and operational data to create a virtual model of Nscale’s operations, marking the first time this lifecycle digital twin architecture has been applied end to end on an active site with both a developer and a tenant.

The pilot, announced on October 21, 2026 at NVIDIA GTC Berlin, focuses on SIN01, the 26MW first building at Start Campus’s 1.2GW SINES Data Campus. By running the DSX-based twin on a live facility, the partners are testing how AI factory decisions on power, cooling and IT can be validated in advance under production-like conditions.

The Immersive Twin is built on NVIDIA Omniverse DSX for visualisation and is linked to Schneider Electric’s real-time simulation engines. ETAP provides power system analysis, while EcoStruxure IT models cooling, with AVEVA software integrating these components into a unified environment. This real-time connection to power and cooling solvers differentiates the system from static, CAD-style digital twins.

Within the model, operators can test hardware configurations, such as placing a Vera Rubin rack in place of a planned Blackwell rack, and immediately see recommendations and deviations on power and thermal performance. This capability is designed to support compressed AI factory timelines, where closely coupled power, cooling and IT leave very limited margin for error and require validation before hardware installation.

The project combines contributions from all three partners: Start Campus supplies facility, engineering and operational data; Nscale provides workload and IT requirements; and Schneider Electric delivers the digital twin technology. Nscale’s chief technology officer has highlighted that the twin gives a clearer view of where full-stack optimisation can take place before physical changes are made at the site.

The SINES Data Campus is planned to host large-scale GPU capacity. Nscale intends to deploy more than 66,000 NVIDIA Rubin GPUs on the campus for Microsoft from late 2027, in SIN02, a second building of up to 200MW. The long expansion runway of the site allows the Immersive Twin concept to be reused across successive builds as the campus grows toward more than 1GW.

Schneider Electric does not yet offer the twin as a generally available product but plans to commercialise the Immersive Twin next year, using the Sines pilot as a template. Each deployment is expected to remain customised to its specific site, with differing building models, while the core technology is standardised for broader use.

Start Campus views the twin as part of its commercial proposition as well as an operational tool. The ability to test infrastructure in a live AI factory environment, Wilson said, enhances the campus’s positioning on scale, sustainability, connectivity and flexibility.