OpenAI and Synopsys partner on an AI model for chip design

OpenAI and chip-design software maker Synopsys announced a multi-year partnership to jointly build GPT-Synopsys, a specialised AI model intended to operate Synopsys’s electronic design automation (EDA) tools directly rather than merely call them through a general-purpose agent. Under the deal, OpenAI will license Synopsys’s EDA tools to develop the model, and the two companies will share revenue and go to market together once it launches; pricing and a release date were not disclosed, though the companies said early technology engagements were already under way with unnamed semiconductor customers.

The companies framed the model as a step beyond existing agentic tools that connect general-purpose AI to design software: rather than operating as an outside user calling EDA tools, GPT-Synopsys is meant to act as what the companies called “a native expert user,” reasoning about chip design and verification and iterating toward a working layout optimised for power, performance and area before handing results to an engineer for review. It is designed to run on OpenAI-hosted infrastructure and integrate with Synopsys’s existing Synopsys.ai and Autopilot platforms.

By helping them build better chips, we can build better AI and bring it to more people.

— Greg Brockman, OpenAI president and co-founder

The deal joined a growing list of efforts to use AI to speed chip design, including Google DeepMind’s AlphaChip, in production use for TPU floorplanning since 2020, and a startup’s unverified claim weeks earlier to have designed and verified a complete accelerator chip in two weeks. Unlike those efforts, Synopsys brought an established EDA business and existing customer base to a partnership explicitly structured around sharing the revenue those customers generate.