at a recent Advancing AI event, AMD shared initial technical specifications for its forthcoming Zen 6-based processor, known as “Venice.”

The “Venice” processor, utilizing a 2nm process, will feature up to 256 Zen 6 cores. This represents a 33% increase compared to the current EPYC ‘Turin’ processor, as AMD aims to enhance compute capabilities optimized for AI applications.

Memory bandwidth is projected to more than double compared to existing EPYC CPUs, reaching 1.6TB/s per socket.”Venice” will also incorporate PCIe Gen6, facilitating quicker communication speeds between cpus and GPUs.

Advancing Performance Boundaries

According to AMD CEO Lisa Su, “”Venice extends our leadership across every dimension that matters in the data center. More performance, better efficiency, and outstanding total cost of ownership.”

Su also stated, “It delivers 70% more compute performance than our current generation leadership Turin CPUs. And to really keep feeding MI400 with data at full speed, at even at rack scale, we have doubled both the GPU and the memory bandwidth and optimized Venice to run at higher speeds. Now we just got Venice back in the labs and it is looking fantastic.”

While “Venice” is the highlight of AMD‘s 2026 platform, Su also provided insights into the company’s plans for 2027.

Su revealed, “We are already deep in the development of our 2027 rack that will push the envelope even further on performance, efficiency and scalability with our next generation Verano CPUs and Instinct MI500-series GPUs, so lots and lots of stuff to come from AMD.”

“Venice extends our leadership across every dimension that matters in the data center…outstanding total cost of ownership.”

This upcoming system builds upon the “helios” rack slated for 2026, which will house 72 MI400 gpus across 18 racks. For 2027, AMD is planning a more extensive configuration consisting of 36 racks.

Based on these details, the next-generation AI rack could perhaps include 144 GPUs. The “Verano” platform is expected to maintain the high bandwidth and performance efficiency established by “Venice,” likely with increased memory and compute capabilities per rack.

While AMD has not officially announced the name of its 2027 rack, sources indicate it is being referred to as a larger-scale continuation of “Helios.”