OpenAI says its models boosted AMD GPU PPA by up to 14%, compressing a year of work into 5 months

mark_k · x · 2026-11-19

OpenAI revealed results from its chip-design work with AMD: its models achieved up to 14% improvements in power, performance, and area (PPA) on heavily optimized GPU designs — compressing roughly a year of engineering into under five months, with some AI-generated improvements already in AMD chips sent for manufacturing.

The post highlights the feedback loop: better AI designs better chips, which enable even better AI.

Related event: OpenAI Models Boost AMD Flagship GPU Design by Up to 14% PPA(2 posts)→

Original post →

More from Infra

Infra channel →