Microsoft and AMD unveil Helios for Azure inference
ryanshrout · x · 2026-07-20
Microsoft and AMD have officially announced Helios, a rackscale platform built on Instinct MI455X GPUs, 6th Gen EPYC "Venice" CPUs, Pensando networking, and ROCm.
Key points:
- Shipments begin in 2H 2026.
- It is positioned for frontier model inference, not training.
- Microsoft will use it for Azure AI services and customer-facing serving workloads.
- Azure ND MI455X v7 VMs are built on Helios for production-scale inference across reasoning, search, and agentic workloads.
- Microsoft also outlined new EPYC Venice VM series: HDv2 for AI data systems and agentic pipelines, and HXv2 for silicon design and technical computing.
The image attached says training will stay on NVIDIA-based systems for now, while Helios becomes Azure’s serving layer.
Related event: AMD and Microsoft Expand Azure AI Partnership with Helios(3 posts)→
More from Companies & People
- Anthropic S-1: Over half of $100K+ customers now use both Claude platform and first-party apps — RebeccaBellan · 2026-10-27
- Marc Andreessen hails Replit CEO's essay: two lessons from computing history on chaos — amasad · 2026-10-27
- Replit CEO on the 'self-driving company': agents to absorb coordination costs by 2030 — amasad · 2026-10-27
- RSNA 2026 Deep Learning Labs return: eight hands-on AI classes over four days in Chicago — PRouzrokh · 2026-10-27
- Anthropic already profitable with 19.1% adjusted operating margin, user notes — PAstynome · 2026-10-27
- Zuckerberg says Apple stopped innovating; Kuprel: it also missed AI smart glasses — Kuprel · 2026-10-27