Working with AMD

OpenAI used its AI models to achieve up to 14% PPA (power, performance, and area) improvement on AMD's flagship GPU in less than five months. According to OpenAI, that's roughly a year's worth of engineering progress.

Changes in production

AMD has already incorporated the improvements into a design sent to production, meaning the AI-driven optimizations are not just lab results but part of a real chip design.

Similar moves in the industry

Anthropic is also applying AI to chip work to optimize cost and performance per dollar. The development shows AI is increasingly taking part in chip design processes.