April 17, 2025
AMD has reached a major milestone in the semiconductor industry with the tape-out of its next-generation EPYC™ processor, codenamed “Venice” — the first high-performance computing (HPC) product in the world to be developed on TSMC’s cutting-edge 2nm (N2) process technology.
This achievement not only demonstrates AMD’s relentless pace of innovation but also highlights the strength of its long-standing collaboration with TSMC, a key partner in co-optimizing advanced design architectures and next-gen manufacturing processes.
“Venice” represents a significant leap in AMD’s data center CPU roadmap. Built on TSMC’s N2 node, the processor promises major gains in performance, power efficiency, and transistor density. This is the first time an HPC CPU has been successfully brought up on this next-generation process, with the official launch expected in 2026.
In parallel, AMD also announced the successful bring-up and validation of its 5th Gen EPYC™ CPUs at TSMC’s new fab in Arizona — reinforcing its commitment to strengthening domestic manufacturing and supply chain resilience in the U.S.
“TSMC has been a crucial partner for many years,” said Lisa Su, AMD Chair and CEO. “Being a lead HPC customer for both the 2nm process and Fab 21 in Arizona is a testament to our close collaboration to push the boundaries of high-performance computing.”
TSMC’s CEO, C.C. Wei, echoed the sentiment: “We are proud to have AMD as a lead partner for our N2 process and Arizona facility. Together, we are driving significant technology scaling — resulting in higher performance, improved efficiency, and better yields for advanced silicon products.”
The move to bring manufacturing to TSMC’s Arizona plant reflects a strategic vision: diversify production, reduce dependency on overseas facilities, and reinforce the U.S. ecosystem in a sector as vital as semiconductors.
With “Venice,” AMD is not only shaping the future of high-performance computing — it’s solidifying its role as a global leader in advanced technology. These new CPUs, born from a cross-continental collaboration, will power the next generation of applications — from artificial intelligence to cloud computing and scientific simulations.
