AMD Helios MI455X 72-GPU rack reaches volume shipment via Supermicro — sovereign-AI buyers get their first non-NVIDIA hyperscale option
Supermicro confirmed early-June volume availability of the AMD Helios 72-GPU rack platform built on Instinct MI455X GPUs and 6th-Gen EPYC Venice CPUs. The rack hits NVL72-class density running ROCm — the first non-NVIDIA hyperscale option to clear the buyer's spec sheet for sovereign-AI and NeoCloud deployments.
The supply-chain piece is the substantive update. AMD has been announcing Helios at CES, Computex, and various partner forums since January; the June volume confirmation moves the platform from press-tour silicon to actual ROC-shippable racks. For the sovereign-AI buyer pool — Nebius, Nscale, Crusoe, European national clouds — that's the moment the procurement calculus changes from "NVIDIA or wait" to "NVIDIA or AMD now."
The strategic gap is still software. NVIDIA's CUDA + NVLink + NeMo + Triton stack remains the production runtime for every frontier lab; ROCm has closed enough of the gap to be viable but not to displace. What Helios changes is the rack-density argument: at 72 GPUs per rack with ROCm portability, AMD is now competitive on the hardware specs that matter most for the next wave of regulatory-driven sovereign deployments. NVIDIA's PC-chip expansion says it's not standing still — but the two-vendor hyperscale market is real for the first time since H100.
Supermicro IR — Supermicro Expands Rack-Scale AI Leadership with AMD Helios Platform → · ServeTheHome — AMD EPYC Venice, Instinct MI455X & Helios Hardware On Display →