AMD recently unveiled its latest Zen architecture roadmap at its Financial Analyst Day 2025 (FAD 2025)， officially confirming the development cycles for two generations of core architectures， Zen 6 and Zen 7， and revealing their key technological directions.
Zen 6 will debut with TSMC’s 2nm process technology， focusing on enhancing performance and efficiency， and is slated for release in 2026.
However， Zen 7 presents even greater anticipation. According to the roadmap， Zen 7 will utilize “Future Node” process technology. While not yet explicitly named， it is expected by industry observers to fall between TSMC’s N2P and N1.8 nodes.
The most significant advancement for Zen 7 lies in the integration of a new Matrix Engine， new AI Data Formats support， and an enhanced AI Pipeline structure.
These upgrades are poised to make Zen 7 the first x86 architecture to feature a dedicated AI acceleration engine， marking the Zen platform’s definitive entry into the AI-native era. This strategic shift signifies AMD’s evolution beyond solely competing on performance and power efficiency， areas where it currently holds a strong advantage， particularly against Intel. Instead， the focus is clearly on AI acceleration as a core tenet for future CPUs.
This strategic direction is not only relevant for the PC and server markets but also indicates AMD’s ambition to play a pivotal role in the burgeoning AI PC ecosystem. Based on current market projections， Zen 7 is anticipated to launch between 2027 and 2028， with the EPYC Verano series data center CPUs expected to be the first products to adopt this new architecture.
