New Delhi: AMD and Meta have jointly unveiled Helios, an open rack-scale reference system aimed at addressing the rising infrastructure demands of large-scale artificial intelligence (AI) and high-performance computing (HPC) workloads. The announcement marks a new step in open data center design through the Open Compute Project (OCP) community.
Meta introduced the Open Rack Wide (ORW) specification — a double-wide rack form factor designed to meet the power, cooling, and serviceability needs of AI-scale data centers. AMD’s Helios aligns with this specification, offering a practical reference for future open infrastructure deployments.
The Helios system integrates AMD Instinct GPUs and EPYC CPUs, targeting up to 1.4 exaFLOPS of FP8 performance and 31 TB of HBM4 memory within a single rack. By adhering to open standards, the design aims to improve interoperability, reduce integration complexity, and accelerate adoption of AI systems across enterprises and hyperscalers.
Forrest Norrod, Executive Vice President and General Manager of AMD’s Data Center Solutions Group, emphasized that open collaboration is central to scaling AI infrastructure efficiently. AMD’s approach with Helios is to make open specifications practically deployable, allowing OEMs, ODMs, and cloud providers to customize and extend systems for specific AI and HPC workloads.
The launch underscores a broader industry shift toward open, scalable data center architectures. As AI workloads grow more complex and resource-intensive, initiatives like Helios reflect an ongoing effort by AMD, Meta, and the OCP community to standardize next-generation infrastructure without locking users into proprietary systems.
Also Read: Ola Electric Enters Rs 1 Trillion BESS Market with Launch of ‘Ola Shakti’
