What changed

OpenAI revealed new details about its collaboration with AMD, where AI models were used to optimize the hardware maker's upcoming flagship GPU and CPU designs. The GPU circuitry saw up to a 14% improvement in power, performance, and area (PPA), while the CPU designs showed unspecified single-digit PPA gains. The AI models were trained to rewrite the code describing the chip’s circuits, achieving results equivalent to condensing a year’s worth of conventional chip design work into under five months.

What this means in practice

The optimization work was conducted in SystemVerilog RTL, a common language for high-end chip design. OpenAI claims its models can reason at the level of an expert microarchitect, enabling further PPA improvements in both GPU and CPU designs. The collaboration is part of a broader trend of AI being used in chip design, with OpenAI previously working with Broadcom on the Jalapeño chip, which was already running workloads in June 2026.

Limitations and availability

The supplied announcement does not specify whether unmodified versions of GPT were used in the design process. OpenAI is also working with Synopsys to train GPT-Synopsis, a model for using chip design tools. The Jalapeño chip is expected to have small initial deployments around late 2026, with broader rollout planned for 2027. The models were tested on chip cores like PicoRV32 and Ibex, with GPT-6.2 Astra showing the best results in optimization scores and cost efficiency.

Source facts for OpenAI's AI Models Help AMD Improve GPU and CPU Designs by 14% PPA: This isn't AI's first time working on chips. The company expected to have small initial Jalapeño deployments around late 2026, and pinned a broader rollout to 2027. The blog post also explains how GPT's chip-design acumen evolved.
Original AINewsTrend explainer based on the cited source. This graphic summarizes source statements; it is not an independent benchmark.

Sources and further reading

  1. tomshardware: OpenAI's AI models boost AMD's upcoming flagship GPU design by up to 14% in PPA