AMD unveiled its Helios rack-scale AI system at Advancing AI 2026, bundling GPUs, CPUs, networking, and software into one deployable unit to challenge Nvidia's data-center lead. Anthropic, OpenAI, Meta, Microsoft, and Oracle have already signed on as customers, with shipments expected to start in the second half of 2026.
AMD is no longer content selling chips — it wants to sell the whole rack. At its Advancing AI 2026 conference on July 23, the company revealed the Helios rack-scale AI system, a fully integrated solution that bundles GPUs, CPUs, networking, and software into a single deployable unit.
What Helios packs into one rack
Each Helios rack combines 72 Instinct MI455X GPUs alongside 18 sixth-generation EPYC "Venice" server CPUs, connected through AMD's Pensando networking hardware and run by its ROCm software stack. The result is a system capable of up to 2.9 exaflops of peak FP4 performance, with 31 TB of HBM4 memory. AMD claims Helios delivers up to 30% more tokens per dollar than Nvidia's Rubin NVL72 rack.
Major AI players are already lining up
Anthropic has committed to deploying up to 2 GW worth of MI450-series GPUs in Helios racks. OpenAI signed on for a 6 GW multigenerational deal. Meta matched that with its own 6 GW commitment, and Microsoft and Oracle round out the customer roster.
Shipments are expected to begin in the second half of 2026 and ramp through 2027.
AMD bets big on server CPUs
AMD projects the server CPU market will surpass $200 billion by 2030. Bank of America has its own forecast, expecting the market to exceed $170 billion, driven largely by the same agentic AI dynamics AMD is targeting. AMD also reported a fourfold increase in AI energy efficiency from 2024 through mid-2026, putting it ahead of schedule on its goal of a 20x improvement by 2030.
Source: Crypto Briefing
Trading involves risk.
Risk Disclaimer
