AMD says 6th Gen EPYC ‘Venice’ CPUs are in production; OEM and cloud deployments targeted for later 2026

AMD has said its 6th Gen EPYC 9006 “Venice” server processors have entered production, with major OEM platforms and deployments by leading cloud providers expected later in 2026. The company has also released a white paper assessing the platform across enterprise, cloud-native, AI and HPC workloads.

The Venice portfolio spans configurations from 8-core processors for edge deployments to 256-core flagship CPUs and rack-scale AI host nodes, according to AMD. The company positions the range as a common software platform for diverse workloads, including agentic AI environments.

AMD’s published comparisons centre on the EPYC 9996. In SPECrate 2026 Integer testing, AMD said the processor delivered 1.2x the per-core performance and 2.24x the platform-level performance of an Nvidia Vera-based platform. In its enterprise and cloud-native testing, AMD reported gains of 2.4x to 3.7x, while claiming a 1.8x to 3.13x advantage over Intel Xeon 6980P in molecular dynamics, materials modelling and weather-forecasting workloads.

For data-center operators managing power-constrained AI capacity, AMD also presented a modeled 100 kW rack comparison. It estimated that EPYC 9996 could provide 3.4x the throughput of a Vera-based platform in that model. This emphasis on performance at a fixed rack-power envelope could be relevant to AI deployments in which CPUs support orchestration, databases, retrieval, networking services and tool execution alongside accelerators.

The performance figures should be treated as vendor claims rather than independently validated operating results. Venice is reported to be in production, but the announced OEM platforms and cloud-provider deployments remain expected later in 2026; AMD did not identify the OEMs or cloud providers, disclose shipment volumes or deployment scale, or provide geographic detail.