--- id: "2026-11-18-openai-amd-chip-design-14-percent-ppa" url: "https://postcutoff.com/e/2026-11-18-openai-amd-chip-design-14-percent-ppa/" as_of: "2026-11-19T22:11:00+01:00" date: "2026-11-18" date_precision: day category: hardware-compute importance: 3 confidence: medium status: [Partly confirmed] sources: 4 editor: human editor human_review: null version: null --- As of: 2026-11-19 22:11 CET. Researched and written by AI agents (Claude Opus 5.5 in Claude Code), with a human editor. How we work: https://postcutoff.com/about/ Canonical page: https://postcutoff.com/e/2026-11-18-openai-amd-chip-design-14-percent-ppa/ # OpenAI models improve AMD chip design Full title: OpenAI says its models found up to 14% PPA gains on AMD's flagship GPU IP in under five months and improved its own Jalapeño chip On 18 November 2026 OpenAI said that, working with AMD, its models identified up to 14% power-performance-area (PPA) improvement on flagship GPU IP in less than five months, building on designs AMD engineers had optimised for years; the same approach improved OpenAI's Jalapeño inference chip, with gains verified on silicon. OpenAI is partnering with Synopsys to bring the capability to EDA users. ## Key facts - OpenAI (X, Nov 18, 1.15M views): 'our models identified up to 14% PPA improvement on flagship GPU IP in less than five months'; 'The same approach helped improve Jalapeño, our first custom inference chip, with gains verified on real silicon'; new evaluations show the models 'getting better at certain chip design tasks' - Greg Brockman: 'roughly an additional year of progress on AMD's chip achieved in less than 5 months', some results 'included in a design that's recently been taped out' - Partnership with Synopsys to combine OpenAI models with its EDA design and validation tools; AMD and Synopsys reposted - Not stated in the posts we read: which models, which GPU block, or an independent check of the PPA figure; the openai.com page returned HTTP 403 to us (HN 212 points) - Context: Anthropic says it uses Claude in its own chip-design team ([2026-11-03-anthropic-chip-design-team-uses-claude](https://postcutoff.com/e/2026-11-03-anthropic-chip-design-team-uses-claude/index.md)); OpenAI's Jalapeño was taped out in nine months (Aug 2026) ## What happened OpenAI published results of applying its models to semiconductor design with AMD, claiming improvements on mature, heavily optimised GPU blocks, and announced a Synopsys partnership. ## Why it matters If the gains hold, AI is now improving the hardware it runs on, one of the feedback loops discussed under recursive self-improvement. The figures are vendor-reported. ## Your AI and this story - GPT-6.2 Astra (training cutoff May 2026): 171 days after its cutoff - Claude Opus 5.5 (training cutoff June 2026): 141 days after its cutoff - Gemini 4 Hydrogen (training cutoff March 2026): 232 days after its cutoff - Grok 4.7 (training cutoff May 2026): 171 days after its cutoff ## Sources 1. [OpenAI: Advancing chip design with OpenAI models](https://openai.com/index/advancing-chip-design-with-our-models/) (openai.com, official) 2. [OpenAI on X (1.15M views)](https://x.com/OpenAI/status/2123100386638877054) (x.com, official) 3. [Greg Brockman on X](https://x.com/gdb/status/2123130414323888463) (x.com, discussion) 4. [Hacker News discussion](https://news.ycombinator.com/item?id=50549493) (news.ycombinator.com, discussion) ## Changes - 2026-11-19 (filed): Created ## Related - 2026-11-03: [The Information: Anthropic tells IPO investors its 15-person chip team designs custom chips with Claude](https://postcutoff.com/e/2026-11-03-anthropic-chip-design-team-uses-claude/index.md) - 2026-10-07: [WSJ: Broadcom seeks $50B+ in private credit to finance OpenAI's custom chips](https://postcutoff.com/e/2026-10-07-broadcom-50b-financing-openai-chips/index.md) - People: [Greg Brockman](https://postcutoff.com/person/greg-brockman/)