AMD Zen 6: EPYC “Venice” Server Chips Arrive With Up to 256 Cores

AMD has officially launched its sixth-generation EPYC server processors, codenamed « Venice, » at its Advancing AI event in San Francisco. Built on TSMC’s 2nm process, the new Zen 6 architecture claims up to 256 cores and more than a 70% jump in performance and power efficiency over the previous generation.

A Chip Built for the AI Compute Boom

AMD CTO Mark Papermaster confirmed that EPYC « Venice » is the first high-performance processor to reach production on TSMC’s 2nm node. With its flagship configuration topping out at 256 Zen 6 cores, the chip marks a 33% jump over the 192 cores of the current EPYC « Turin. » AMD claims an overall performance and efficiency improvement of more than 70% over the Zen 5-based generation.

On the platform side, « Venice » introduces the new SP7 socket, with 16-channel memory delivering up to 1.6 TB/s of bandwidth, along with a move to PCIe Gen 6 for very high-speed data transfers. These specs are aimed squarely at AI servers, where memory bandwidth is often the main bottleneck.

Built for AMD’s « Helios » AI Servers

These new EPYC « Venice » chips will primarily power AMD’s Helios AI servers, designed to compete directly with Nvidia’s offerings in AI training and inference infrastructure. With more cores, more memory channels, and faster connectivity than the Turin generation, AMD is positioning itself as a credible alternative to Nvidia’s current dominance of the AI chip market.

Why It Matters

This launch doesn’t directly affect consumers — EPYC « Venice » targets data centers and cloud providers. But its effects should be felt indirectly, through potentially lower inference costs for the AI services you use every day (assistants, image generation, search), as cloud providers roll out this more efficient generation of servers.

The launch also confirms AMD’s strategy: a tight annual release cadence for its server chips, with a desktop version of the Zen 6 architecture expected to follow, potentially as soon as later this year.

It remains to be seen how Intel and Nvidia will respond to this push into the server CPU market for AI — a fast-growing segment where every performance gain translates directly into economies of scale for cloud giants.

What do you think of AMD’s push against Nvidia in the AI chip market? Share your thoughts in the comments and follow Wanda-techs for all the latest hardware news.