Siemens, NVIDIA, and Fluence unveil reference architecture for 136 MW AI data center power systems

Siemens has partnered with NVIDIA and Fluence, while incorporating nVent-aligned design considerations, to create an NVIDIA DSX Vera Rubin-aligned reference design that transforms NVIDIA’s AI factory concept into a practical electrical, power, and controls architecture for hyperscale, colocation, and specialized cloud data centers.

NVL72 reference design built for next-generation AI infrastructure

The rapid rise of AI factories is reshaping data center requirements, with platforms such as NVIDIA Vera Rubin NVL72 placing greater demands on power delivery, cooling systems, efficiency, and reliability. Operators are increasingly challenged to balance site selection, grid access, investment efficiency, and speed to revenue while adopting new technologies.

Siemens’ reference design supports a facility capacity of 136 MW, including a 100 MW IT load. It provides a complete power and controls framework, covering everything from a 34.5 kV utility connection through medium-voltage distribution and modular low-voltage power systems to the rack level.

Based on a Tier III concurrent maintainability approach, the design ensures maintenance can be performed on individual components without affecting IT operations. Its modular building-block architecture aligns with NVIDIA DSX Vera Rubin deployment units, allowing capacity to expand from tens of megawatts to hundreds of megawatts without major redesign.

To support NVIDIA workloads and system architectures, the design incorporates nVent-aligned electrical parameters. A future supplement is expected to add advanced thermal management capabilities.

The reference design enables AI factory operators to deploy ultra-dense AI environments with greater speed, predictability, and reduced risk while maintaining flexibility for future technology upgrades and changing energy needs. It also supports DSX MaxLPS, helping maximize computing performance and token generation within fixed power budgets.

Fluence contributes battery energy storage technology that strengthens flexibility and resilience for AI factories operating in constrained power environments. According to Jeff Monday, Fluence Chief Growth Officer, the Smartstack platform delivers critical grid-support functions including voltage and frequency ride-through, black start capability, demand response, and AI load smoothing.

The design is integrated with a centralized Integrated Data Center Management Suite that provides a unified operational view across power, cooling, and compute infrastructure.

Through its portfolio of power distribution systems, automation technologies, energy management software, connected hardware, AI applications, cloud platforms, and digital services, Siemens helps data center operators scale confidently and meet the growing infrastructure requirements of AI-driven computing.