AMD invests over $10B in Taiwan for advanced chip packaging with TSMC

AMD announced on May 21, 2026, that it will invest over $10 billion (NT$315.58 billion) in Taiwan's semiconductor ecosystem, focusing on advanced chip packaging technology developed with TSMC. The multi-year investment, covering 2026 through 2029, targets EFB 2.5D packaging to support AMD's sixth-generation EPYC server CPUs (codenamed Venice) and the Helios rack-scale AI platform with Instinct MI450X GPUs, positioning AMD against Nvidia in the AI server market.

Photo: Grujoungarot / Wikimedia Commons / CC BY-SA 4.0 https://creativecommons.org/licenses/by-sa/4.0 AMD invests over $10B in Taiwan for advanced chip packaging with TSMC Lisa Su's multi-year bet on next-generation 2.5D packaging technology targets AI infrastructure bottlenecks and positions AMD squarely against Nvidia in the server market AMD is writing a very large check to Taiwan. The company announced on May 21, 2026 that it will pour more than $10 billion, equivalent to over NT$315.58 billion, into Taiwan’s semiconductor ecosystem over the next several years, with a specific focus on advanced chip packaging technology developed alongside TSMC and a roster of local manufacturing partners. This is not a vague commitment to future R&D. The investment is structured to secure production capacity from 2026 through 2029, targeting a relatively new packaging approach called Elevated Fan-Out Bridge, or EFB, 2.5D packaging technology. What AMD is actually buying EFB-based 2.5D packaging is AMD’s chosen architecture for solving that connectivity problem. It improves interconnect bandwidth, meaning chips can pass data between each other faster, while also consuming less energy in the process. Both matter enormously when you are building AI infrastructure at the scale AMD is targeting. The specific products this investment is designed to support include AMD’s sixth-generation EPYC server CPUs, codenamed Venice, which are transitioning to TSMC’s 2nm manufacturing process. It also underpins the Helios rack-scale AI platform, which features AMD’s Instinct MI450X GPUs. AMD has flagged multi-gigawatt deployments of that platform for the second half of 2026. Partners named in the effort include ASE/SPIL, PTI, and Sanmina, all established players in semiconductor assembly and packaging. The Nvidia problem AMD is trying to solve Nvidia has leaned heavily on TSMC’s CoWoS, or Chip-on-Wafer-on-Substrate, technology, and the limited capacity for that process has been a genuine constraint on the broader AI chip market. AMD’s EFB investment is a calculated attempt to build capacity in a comparable but distinct packaging architecture. CEO Lisa Su framed the announcement as a long-term investment in Taiwan’s semiconductor ecosystem. Taiwan, geopolitics, and the Arizona hedge Lisa Su noted that the Taiwan investment operates alongside AMD’s use of TSMC’s Arizona facilities, which are part of TSMC’s US expansion program receiving support under the CHIPS Act framework. Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy https://cryptobriefing.com/editorial-policy/ .