AMD is chucking more than $10 billion at Taiwan as it prepares Helios racks for the next AI feeding frenzy.
The move comes as the company nears deployment of sixth-generation EPYC chips and Instinct MI450X GPUs in its Helios AI racks.
AMD says it is ready to meet the rising demand for AI by putting serious money into Taiwan and its regional supply chain. Those partners will help supply Helios, one of the fastest AI racks the outfit has cooked up so far.
The AI-focused rack houses AMD’s sixth-generation EPYC CPUs, codenamed Venice, and Instinct MI450X AI GPUs.
The company said its more than $10 billion investment across Taiwan will expand strategic partnerships and scale advanced packaging for AI infrastructure. AMD is pushing industry-leading EFB-based 2.5D packaging to boost interconnect bandwidth and efficiency in Venice CPUs.
The Helios rack-scale platform, packed with Venice and AMD Instinct MI450X GPUs, is on track for multi-gigawatt deployments beginning in the second half of 2026.
AMD is working with Taiwan-based ODMs, including Sanmina, Wiwynn, Wistron and Inventec. It is joined by SPIL, PTI, Unimicron, AIC, Nan Ya PCB and Kinsus, which will build and ship the Helios AI racks.
This is AMD’s bid to lay the tracks for its agentic AI roadmap before the AI gravy train runs out of rails. The company said the plan is aimed at next-generation AI infrastructure rather than another shiny deck for investors.
The efforts build on AMD’s chiplet work, high-bandwidth memory integration, 3D hybrid bonding and rack-scale system design.
AMD chair and chief executive Dr Lisa Su said customers were scaling AI infrastructure quickly as demand for compute continues to rise.
“As AI adoption accelerates, our global customers are rapidly scaling AI infrastructure to meet growing compute demand. By combining AMD leadership in high-performance computing with the Taiwan ecosystem and our strategic global partners, we are enabling integrated, rack-scale AI infrastructure that helps customers accelerate deployment of next-generation AI systems.”
The company has achieved a milestone with PTI by qualifying what it calls the industry’s first 2.5D panel-based EFB interconnect. That technology supports high-bandwidth interconnect at scale, which should enable customers to deploy more efficient AI systems while improving economics.
Together, AMD says these moves reinforce its leadership in delivering high-performance AI infrastructure at scale.
