NuAxiSemiconductor trade intelligence

AMD Launches 6th-Gen EPYC Venice, Claims 3.75 Times Nvidia Performance

What happened

AMD officially launched its sixth-generation EPYC server CPUs (codenamed Venice) based on the Zen 6 microarchitecture on Nov. 10. The processors are now commercially available, with major server manufacturers beginning shipments of Venice-based systems, and hyperscale cloud providers including Microsoft Azure and Oracle Cloud Infrastructure (OCI) planning to make VM instances powered by the processors available to general customers soon. AMD divided the Venice lineup into three categories: GPU host nodes, multi-agent AI workloads, and general-purpose enterprise computing. In the GPU host-node category, a high-frequency 96-core model delivers 1.8 to 1.9 times the AI inference output throughput on Nvidia B200 GPUs versus fifth-generation EPYC processors. The EPYC 9996 scored 5,470 points in the SPECrate 2017 integer benchmark versus 1,459 points for a system powered by Nvidia's Vera CPU (both dual-socket), a performance advantage of up to 3.75 times; versus Intel's sixth-generation Xeon Granite Rapids 6980P it delivered 2.74 times higher performance by geometric mean of multiple benchmarks, including 3.45 times higher planning and orchestration performance in TPCx-AI, 2.83 times higher web gateway request throughput measured with WRK and NGINX, 2.65 times higher performance across enterprise tools and network workloads, and 2.5 times higher ephemeral tool execution performance. The EPYC 9996 features 256 cores, a 600-watt default CPU power (DCP) rating and is priced at $14,904, versus Intel's Xeon 6980P with 128 cores, a 500-watt TDP and a $13,955 price. AMD also announced the sixth-generation EPYC Embedded 9006 Series SP7 CPUs for the embedded computing market. The source states the launch aims to expand market share by offering better price-performance efficiency than competing products.