OpenAI and Synopsys signed a multi-year deal to build GPT-Synopsys, a specialized “frontier intelligence” model meant to run Synopsys electronic design automation tools the way expert chip engineers do — not just chat about design, but operate the loop.

According to Synopsys’ Sept. 30, 2026 announcement, the companies will act as preferred partners on R&D and a joint go-to-market, with a revenue-sharing arrangement. OpenAI will license Synopsys EDA software to develop the model; GPT-Synopsys is slated to run on OpenAI-hosted infrastructure and plug into Synopsys.ai and Synopsys Autopilot. Synopsys says early technology engagements with leading semiconductor customers are already underway. There is no general-availability date and no disclosed dollar figure for OpenAI’s side of the partnership.

What changes in the design loop

Synopsys frames today’s agentic setups as general-purpose models bolted onto EDA tools. GPT-Synopsys is the next step the companies want: a frontier model that learns to drive those tools, read their outputs, and iterate on power, performance, and area — plus timing and verification closure — with engineers still reviewing verified outcomes. The pitch is more design alternatives, faster path to working silicon, without dumping first-time-right discipline.

Synopsys CEO Sassine Ghazi cast the deal as expanding access to Synopsys’ “ground-truth” engineering tools while keeping manufacturing rigor. OpenAI president and co-founder Greg Brockman said the company is using its most advanced technology on the systems that power AI — and, with Synopsys, bringing that work into chip design so better chips can feed better AI.

What’s shipping vs what’s promised

The release describes a joint service that bundles compute, model, and licenses, with enterprise controls and a claim that customer design data is not used to train the model. Those are company assertions. Commercially, Synopsys is careful: availability and timing sit in forward-looking language, and the only concrete customer status is early engagements, not a broad GA launch.

Same-day context: Amazon IP

Separately on the same day, Synopsys announced a multi-year, $1B+ silicon IP agreement with Amazon for custom chips and AWS infrastructure — a different deal, with dollars attached. That pact is backdrop for Synopsys’ silicon-and-systems push, not the GPT-Synopsys lead.

Stakes: If frontier models become native EDA operators under a shared-revenue GTM, chip houses may buy design capacity as a model-plus-tools service — but until Synopsys and OpenAI price the stack and exit “early engagement,” the EDA loop shift remains partnership theater with real product still in the hangar.