AMD to Invest Over $10 Billion in Taiwan's Ecosystem to Accelerate AI Infrastructure

U.S. chip manufacturer Advanced Micro Devices (AMD) announced today that it will invest over $10 billion in Taiwan's industrial ecosystem to meet the growing demand for AI infrastructure, expand strategic partnerships, and enhance advanced packaging manufacturing capacity for next-generation AI infrastructure.

In a press release, AMD Chair and CEO Lisa Su stated, "As AI applications continue to accelerate in adoption, AMD's global customers are rapidly expanding their AI infrastructure to meet the ever-increasing computational demands. By combining AMD's high-performance computing with Taiwan's industrial ecosystem and global strategic partners, we are realizing integrated, rack-scale AI infrastructure to help customers accelerate the deployment of next-generation AI systems."

AMD noted that based on its deep partnerships with the industry ecosystem and its strengths in chiplet architecture, High-Bandwidth Memory (HBM) integration, 3D hybrid bonding, and next-generation AI infrastructure rack-scale system design, AMD is continuously advancing chip, packaging, and manufacturing technologies to achieve higher performance, better efficiency, and accelerate AI system deployment.

AMD stated that the investment plan announced today demonstrates its extension of leadership through strategic partnerships to drive the chip, packaging, and manufacturing innovations required for next-generation AI infrastructure.

Among these efforts, AMD is collaborating with Taiwan's ASE, SPIL, and other industry partners to jointly develop and validate next-generation wafer-based 2.5D bridge interconnect technology.

AMD pointed out that the EFB (Elevated Fanout Bridge) architecture significantly increases interconnect bandwidth and improves power efficiency, providing strong support for the "Venice" Central Processing Unit (CPU) to deliver higher performance-per-watt under practical power and thermal constraints.

AMD also announced a significant milestone with Powertech Technology Inc. (PTI), successfully validating 2.5D panel-level EFB interconnect technology. This technology supports large-scale, high-bandwidth interconnects, enabling customers to deploy more efficient AI systems while improving overall economic benefits.

AMD emphasized that these technological advancements collectively solidify its leadership in the large-scale deployment of high-performance AI infrastructure. By combining chip innovation with a robust global industry ecosystem, AMD is helping customers accelerate the deployment of next-generation AI systems.

Furthermore, AMD and its ecosystem partners are leveraging these innovations to support the deployment of the AMD Helios rack-scale platform in the second half of 2026, marking a significant step towards production-ready AI infrastructure.

AMD noted that ODM partners such as Sanmina, Wiwynn, Wistron, and Inventec are helping to build systems based on AMD Helios. This system features the AMD Instinct MI450X Graphics Processing Unit (GPU), the 6th Gen AMD EPYC CPU, advanced networking solutions, and the AMD ROCm open software stack, facilitating a smooth transition from platform design to large-scale mass production.

AMD stated that the Helios platform delivers disruptive AI performance through breakthroughs in compute, interconnect bandwidth, memory capacity, and system-level integration, allowing customers to run larger, more complex AI workloads faster while optimizing power and efficiency.