Schneider Electric, AVEVA, Nscale, and Start Campus, announced one of the first pilots of an end-to-end digital twin developed using the NVIDIA Omniverse DSX Blueprint, part of NVIDIA DSX Sim.
The Immersive Twin pilot marks a significant step toward bringing Schneider Electric’s digital twin technology, powered by AVEVA software, into real-world AI infrastructure, allowing Nscale to accelerate token production and optimise token efficiency.
The deployment is focused on Nscale’s operations within the SIN01 data centre on the 1.2 GW SINES Data Campus in Portugal, developed by Start Campus. As one of Europe’s largest and most sustainable AI factories, the SINES Data Campus is designed to support successive generations of AI infrastructure. It will soon become home to more than 66,000 NVIDIA Rubin GPUs, representing one of the largest AI deployments in the region. By combining engineering and operational data with advanced power and cooling simulations and workload requirements, the new Schneider Electric and AVEVA Immersive Twin provides a unified view of data centre performance to help identify opportunities to optimise infrastructure before changes are implemented on-site.
Using the digital twin, Nscale expects to achieve faster time to token, greater token efficiency, enhanced operations, increased flexibility, and reduced risk when updating or modifying its technology stack.
“Working with Start Campus, Schneider Electric and NVIDIA, we’re using digital twins to model how AI data centres operate in the real world — from power and cooling to token delivery,” said David Power, Chief Technology Officer at Nscale. “This gives us a clearer view of where full-stack optimisations can be realised before changes are made on the ground. It’s about using intelligence and simulation to build and operate AI infrastructure more efficiently, and sustainably at scale.”
Complete end-to-end experience and collaboration
The successful pilot reflects the strength of a collaborative, end-to-end approach, bringing together complementary expertise from across the AI ecosystem, including energy management, engineering, simulation and AI infrastructure operations to create a more complete view of data centre performance.
The Immersive Twin pilot supports the full lifecycle of AI factories—from design through operations—by building an OpenUSD-based digital twin that connects Schneider Electric, ETAP, AVEVA, and third-party technologies. It enables rapid “what-if” scenario testing across power, cooling, and process behaviour to validate design decisions, assess upgrade impacts, reduce procurement risk, and continuously optimise performance by linking IT and OT data. In parallel, it advances interoperability through simulation-ready asset files, accelerating seamless integration into digital twins.
Start Campus provides the real-world infrastructure, engineering expertise and operational environment to demonstrate how digital twin technology can transform the design, development and operation of AI data centres. Bringing this data, technology and expertise together enables Nscale to use the digital twin to model and test “what-if scenarios” for new configurations or technologies before implementation, speeding up deployment times while significantly reducing risk.
“We are proud to have joined forces with Schneider Electric, NVIDIA, Nscale and Start Campus to develop and deploy this unified digital twin, purpose-built for AI factories,” said Caspar Herzberg, CEOof AVEVA. “By bringing together power, cooling and infrastructure in one environment, operators can simulate upgrades, expansions and greenfield projects while continuously optimising physical operations as conditions change, improving both ‘time to token’ and overall efficiency. Built on proven AVEVA technologies, including CONNECT, PI System, AVEVA Process Simulation, AVEVA Unified Engineering and AVEVA Asset Information Management, the digital twin supports both simulation and real-time visibility into asset performance in a single pane of glass, enabled by Viewport, the visualisation application within NVIDIA Omniverse.”
“AI infrastructure needs to evolve as quickly as the technology it supports,” added Robert Dunn, Chief Executive Officer at Start Campus. “Our collaboration with Schneider Electric, Nscale and NVIDIA has taught us that operators and customers design, develop and operate data centres best when they work from the same model. By using digital twin technology to simulate how a jump in rack density moves through the power train and the liquid cooling loop before it becomes a physical reality, we can test new hardware generations, agree the configuration with our customer and build it once. That accelerates deployment, improves efficiency and maximises the value of every megawatt. This is an important step towards delivering resilient AI infrastructure at scale.”
Building on existing partnerships
This latest pilot expands Schneider Electric, AVEVA and ETAP’s work with NVIDIA. Last year, the three companies joined the Alliance for OpenUSD (AOUSD) to accelerate digital twin and AI factory development through NVIDIA Omniverse DSX Blueprint.
Additionally, the companies to collaborate on reference architectures, integrated infrastructure and software solutions that will power and cool the next generation of AI factories, and leveraging NVIDIA Omniverse libraries to simulate and improve energy efficiency in real-world environments.
Discussion about this post