cd /news/ai-chips/amd-claims-epyc-venice-beats-nvidia-… · home topics ai-chips article
[ARTICLE · art-134708] src=techpowerup.com ↗ pub= topic=ai-chips verified=true sentiment=· neutral

AMD Claims EPYC "Venice" Beats NVIDIA Vera by 2.24x in New White Paper

AMD published a white paper claiming its 256-core 6th Gen EPYC 9006 "Venice" server CPU delivers 2.24x the platform-level performance of NVIDIA's 88-core Vera in SPECrate 2026 Integer, with a 96-core high-frequency Venice part claimed 1.2x faster per core. AMD also reports 2.4x to 3.7x gains over Intel's Xeon 6980P in server-side Java, OpenSSL, MongoDB, Redis, NGINX and transaction processing, and up to 3.13x in HPC tests like GROMACS, though the figures are first-party estimates and neither company can publish official SPEC results yet. Tom's Hardware noted the comparison used GCC 16.1 on AMD against GCC 15.2 for NVIDIA and pitted a 512-thread flagship against Vera's 176 threads, so it is not like-for-like; AMD says Venice is in production with OEM platforms on track to launch and cloud deployments starting later this year.

by read1 min views2 publishedSep 19, 2026

AMD has published a white paper on its 6th Gen EPYC 9006 "Venice" server CPUs, and it takes direct aim at NVIDIA's Vera. These are first-party figures from AMD, so it is important to be cautious about their validity. In SPECrate 2026 Integer, AMD says the 256-core EPYC 9996 delivers 2.24 times the platform-level performance of an 88-core Vera system, while a 96-core, high-frequency Venice chip is claimed to be 1.2 times faster per core. Against Intel's Xeon 6980P, AMD reports gains of 2.4x to 3.7x in server-side Java, OpenSSL, MongoDB, Redis, NGINX and transaction processing, and up to 3.13x in HPC tests like GROMACS. Neither AMD nor NVIDIA can publish official SPEC results yet, so all of it is estimated.

The Vera figures come from NVIDIA's own white paper, and the 96-core part is an EPYC 9996 with several of its cores turned off. The per-subtest results also used GCC 16.1 on the AMD side against GCC 15.2 for NVIDIA, which Tom's Hardware notes is not best practice. The 2.24x figure compares a 512-thread flagship with Vera's 176 threads, so it is not a like-for-like matchup. AMD says Venice is in production, with major OEM platforms on track to launch and cloud providers starting deployments later this year. It's likely we will only get a real answer once independent reviewers get both chips on the same test bench.

── more in #ai-chips 4 stories · sorted by recency
── more on @amd 3 stories trending now
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/amd-claims-epyc-veni…] indexed:0 read:1min 2026-09-19 ·