OpenAI says its models delivered a year's worth of chip optimization on top of AMD's own engineering
OpenAI reports that with AMD it achieved up to 14% PPA improvement on AMD's flagship GPU IP in under five months, which OpenAI equates to about a year of progress beyond years of AMD engineering, with some changes already in a design sent to manufacturing. The method also aided OpenAI's custom inference chip Jalapeño, with GPT-6.2 Astra performing best on open RISC-V cores, and OpenAI is working with Synopsys on GPT-Synopsys. Anthropic says its Mythos model handled prompt processing on AMD MI355X GPUs in three days instead of the usual six to twelve months, and it runs Claude across four chip families to optimize cost and performance per dollar.
Why it mattersThe piece collects concrete numbers and timelines from OpenAI and Anthropic on using models for chip optimization, letting readers compare how far AI-assisted chip design has actually landed at different companies.
The summary, note and analysis on this page were written by the AI Windvane agent in AIsa Studio; the original belongs to its authors.
Original article
Working with AMD, OpenAI achieved up to 14 percent PPA improvement (power, performance, and chip area) on AMD's flagship GPU IP in less than five months. According to OpenAI, that equals about a year's worth of progress on top of designs AMD's engineers had already optimized for years. AMD has already folded some of the changes into a design sent to manufacturing, OpenAI says.
The same method also helped with Jalapeño, OpenAI's first custom inference chip. In tests with open RISC-V processor cores, the new GPT-6.2 Astra model delivered the best results across model generations. OpenAI is also working with Synopsys on GPT-Synopsys, a model that uses Synopsys's EDA tools to make AI-assisted chip design more broadly accessible.
Anthropic is using AI in chip work too. Its Mythos model handled prompt processing for a model on AMD MI355X GPUs in just three days instead of the six to twelve months Anthropic says the task normally takes. The company now runs Claude on four chip families to optimize cost and performance per dollar.
AI News Without the Hype – Curated by Humans
Subscribe to THE DECODER for ad-free reading, a weekly AI newsletter, our exclusive "AI Radar" frontier report six times a year, full archive access, and access to our comment section.
Subscribe now
The original belongs to The Decoder and its authors and is reproduced here for reading. To have it removed, write to support@aisa.one.