OpenAI-assisted chip design reportedly improved AMD’s next flagship GPU by 14% PPA
AMD has begun using OpenAI’s GPT models to assist with chip development, according to Tom’s Hardware. The company applied the technology to an upcoming flagship GPU and its new CPUs.
For the GPU, the report claims that AI-assisted design work produced a 14% improvement in PPA, a metric commonly used to assess performance, power consumption, and chip area. The headline also indicates that GPT models contributed to improvements in AMD’s forthcoming CPUs, although the available report does not provide specific methodology or performance results for those processors.
The development suggests that generative AI models are being used beyond software applications as engineering assistants in semiconductor design, where small gains across power, performance, and area can affect a product’s capabilities, efficiency, and manufacturability.