AMD to pour $10B+ into Taiwan AI supply chain

AMD announced on May 21, 2026 that it will invest more than $10 billion across Taiwan’s ecosystem to expand strategic partnerships and scale advanced packaging manufacturing for AI infrastructure. The company said the effort is aimed at improving silicon, packaging, and manufacturing capabilities so AI systems can be deployed faster and run more efficiently.

A central focus is AMD’s work with Taiwan-based packaging and testing firms ASE and SPIL on next-generation wafer-based 2.5D bridge interconnect technology, known as EFB, which is intended to increase bandwidth and improve power efficiency for its 6th Gen EPYC processors codenamed Venice. AMD also said it has qualified the industry’s first 2.5D panel-based EFB interconnect with Powertech Technology, a milestone it says supports higher-scale, more economical AI system production.

The investment also supports AMD’s Helios rack-scale platform, which combines Venice CPUs, Instinct MI450X GPUs, networking, and AMD’s ROCm software stack, with partner help from Sanmina, Wiwynn, Wistron, and Inventec. AMD said Helios-based systems are on track for multi-gigawatt deployments beginning in the second half of 2026, underscoring that the company is trying to move beyond chip design into the broader AI infrastructure stack.