AMD Embraces 2nm Technology

AMD has officially confirmed that its next-generation EPYC server processors, codenamed Venice, have entered the production phase. These chips utilize TSMC’s sophisticated 2nm manufacturing process, with plans to expand production to TSMC’s facility in Arizona in the future. Venice is built upon the upcoming Zen 6 CPU architecture, which is expected to underpin the next generation of desktop processors, internally referred to as Olympic Ridge and potentially marketed under the Ryzen 10000 series branding.


The transition to TSMC’s N2 node marks a significant strategic pivot for AMD. Having utilized the N5 node extensively for its current lineup, including the Ryzen 9000 series, the company appears to be largely bypassing the N3 node in favor of this more advanced technology. While commercial 2nm chips are currently unavailable, Apple is widely expected to debut such silicon in upcoming hardware. AMD’s aggressive adoption of this process suggests a major shift in its manufacturing strategy.


Performance Expectations and Timeline

One of the primary benefits of moving to N2 silicon is the potential for increased density. Engineers expect to see a jump to 12 cores per chiplet (Core Complex Die), a significant increase from the eight cores found in current models. This development could pave the way for:

  • Gaming-optimized processors integrated with 3D V-Cache.
  • High-performance dual-die configurations featuring up to 24 cores for intensive multithreaded tasks.

Regarding market availability, AMD anticipates the Zen 6 architecture will debut in late 2026. Industry experts suggest that the initial rollout will likely focus on EPYC server components, with consumer desktop models following in early 2027.


Intel’s Strategic Vision

While AMD pushes forward with 2nm production, Intel continues to refine its own foundry strategy. Currently, Intel is deploying its 18A node for Panther Lake mobile CPUs and is actively developing its 14A node. In a notable development, Intel leadership recently hinted at future 10A and 7A nodes during the J.P. Morgan annual tech conference.


«People don't go to you for just one node. They're looking for the roadmap for the future,» stated Intel leadership regarding the importance of long-term planning for their foundry customers. Although these 10A and 7A nodes remain speculative—especially given that 14A production is not expected until 2029—the public mention of such advanced milestones represents a significant shift in corporate confidence and long-term manufacturing commitment.


The State of Moore’s Law

While the original definition of Moore’s Law—doubling transistor density while halving costs every two years—may have reached its limits, these developments prove that the semiconductor industry is far from stagnant. Through consistent advancements in node technology from both AMD and Intel, the industry continues to push the boundaries of what is possible in silicon performance.