Schneider Electric and AMD have released a reference design for the AMD Helios rack-scale platform, which they describe as the first of its kind for Helios.
The design is intended to give data centre operators and architects a tested layout for deploying high-density AI infrastructure, covering power, cooling, IT space and lifecycle software. It is the first result of a collaboration between the companies focused on large AI installations.
The reference design supports racks drawing up to 246 kW and modular AI clusters with IT loads of up to 10.4 MW. It is aimed at both new AI facilities and retrofit projects in existing high-density sites.
Helios is built around AMD hardware, including Instinct MI455X GPUs, sixth-generation EPYC CPUs and Pensando Vulcano network interface cards, alongside the ROCm software stack. Schneider Electric's contribution focuses on power systems, cooling equipment and digital infrastructure tools used to plan and run data centres.
Reference designs are becoming more important as AI workloads place heavier demands on data centre infrastructure. Operators now need to manage higher rack densities, more complex thermal conditions and tighter integration between compute systems and building infrastructure.
The blueprint includes facility power planning, cooling, IT room layout and lifecycle software guidance. By pre-validating those areas, the design is intended to shorten planning and reduce the likelihood of deployment problems.
Cooling focus
A central feature is liquid cooling. The design uses CDU-based and hybrid air-and-liquid cooling systems from Motivair, Schneider Electric's cooling business. According to the companies, those systems can remove up to 84% of heat.
They also said the design can achieve a power usage effectiveness rating of about 1.12 at full load. PUE is a common measure of data centre efficiency, comparing total facility energy use with the energy consumed by IT equipment.
Beyond physical infrastructure, the design uses ETAP and EcoStruxure IT Design computational fluid dynamics simulation tools for electrical and thermal modelling. The framework also includes an electrical digital twin and support from AVEVA's Unified Operations Centre for monitoring and operational visibility.
That reflects a wider shift in AI data centre development, as software used to model power and cooling becomes part of the early design process rather than a later operational layer. With high-density AI systems, infrastructure constraints can shape site design as much as the choice of processors or networking equipment.
Schneider Electric said the reference design has been validated to ANSI standards for deployments in the United States. The companies plan to extend the framework to support IEC standards for wider international use.
For AMD, the launch positions Helios as a rack-scale system rather than simply a collection of chips and interconnects. The company has been pursuing a broader data centre platform strategy as customers look for integrated AI systems that can be installed more quickly and scaled in larger blocks.
For Schneider Electric, the move highlights the growing commercial importance of infrastructure design in the AI buildout. As demand for AI computing rises, suppliers of electrical equipment, cooling systems and monitoring software are seeking a larger role in the earliest stages of project planning.
Manish Kumar, executive vice president of Secure Power & Data Centres at Schneider Electric, said organisations now want tested designs that connect planning and deployment. "Today organizations require comprehensive, AI-ready reference designs that can take them from planning to deployment faster and with less risk," Kumar said. "Through our collaboration with AMD, we're delivering an engineering-backed reference design that bridges the gap between advanced AI compute platforms, energy tech, and real-world data center implementation, enabling customers to deploy scalable, high-density AI environments with greater confidence, efficiency, and speed."
AMD described the release as part of a shift towards larger AI production environments that require infrastructure and computing hardware to be designed together. "AI infrastructure is rapidly moving to full-scale AI factories, and that requires compute, networking, power and cooling to be designed together from the start," said Forrest Norrod, executive vice president and general manager of the Data Centre Solutions Business Group at AMD. "AMD Helios provides an open, rack-scale architecture built to deliver the performance, efficiency and flexibility required for next-generation AI workloads. By working with Schneider Electric to create a validated reference design, we are giving customers a practical blueprint to accelerate high-density AI deployments, reduce integration risk and scale with greater confidence and efficiency."