OpenAI and Synopsys plan a chip-design model that can use EDA tools

OpenAI and Synopsys announced a multi-year agreement on September 30 to develop GPT-Synopsys, a specialized model intended to operate Synopsys electronic design automation (EDA) tools. OpenAI will license those tools for development. The companies also plan a joint commercial offering, with revenue sharing and early technology engagements already underway with semiconductor customers.

The release describes a development program and an intended service. It gives no public launch date, access procedure, price, model specification or measured GPT-Synopsys result. For a chip team deciding what it can deploy today, that boundary matters as much as the partnership itself.

The big change

• What changed: OpenAI and Synopsys have agreed to develop a specialized model that operates chip-design tools. The proposed service would bundle the model, computing capacity and Synopsys licenses, according to the companies.
• Why it matters: Engineers would be able to assign design objectives while agents operate EDA tools, inspect their output and propose iterations for human review, if the planned workflow works as described. That would move some tool operation into the model's job; the September 30 release does not establish a faster or better finished chip.
• What to watch: Customer access terms, supported workflows and measured design results will determine whether this becomes a useful engineering service. The companies report early engagements, but have not published those details or an availability date for GPT-Synopsys.

The workflow the companies describe

Synopsys says the intended model will reason about design and verification and directly operate its EDA software. An engineer would delegate an objective such as power, performance and area optimization, timing closure or verification closure. Agents would run tools, interpret their outputs, change the design and repeat the cycle before presenting a result for engineer review. The release says GPT-Synopsys is planned to run on OpenAI-hosted infrastructure, interoperate with customer agent harnesses and integrate with Synopsys.ai and the Synopsys Autopilot platform.

That sequence is a useful way to understand the ambition, but the announcement contains no public demonstration record or benchmark for GPT-Synopsys. It does not identify the underlying OpenAI model, tool interfaces, supported design flows, engineering prerequisites or validation criteria. “Verified outcomes for engineer review” is the companies' description of the intended process; it is not evidence that a customer has reached silicon signoff with this model.

The September 30 release says the planned service will bundle compute, model access and tool licenses. It also describes security and data controls, including encryption, configurable retention and an assurance that customer data is not used to train the model. Those are vendor statements about the offering, not independently tested properties of a publicly documented deployment. A prospective customer still needs product terms and technical documentation to assess data handling, integration and cost.

What teams can use now

GPT-Synopsys is distinct from Synopsys's existing AI tools. Synopsys describes five Synopsys.ai Copilot assistants and advisors as commercially available, including tools for knowledge retrieval, scripts, RTL code and formal verification. The company reports productivity gains for those products, but those figures are not measurements of GPT-Synopsys.

Synopsys also announced AgentEngineer solutions and its Autopilot platform on September 28. It said more than 50 engagements were underway for those products and that availability was planned for the end of 2026. That timetable applies to the September 28 portfolio, not to GPT-Synopsys. The new partnership release says GPT-Synopsys will integrate with Autopilot but gives no date for the specialized model.

The practical next step for teams interested specifically in GPT-Synopsys is to seek access, licensing and deployment details from the companies. The public announcement cannot yet support a procurement comparison or a design-performance claim. Synopsys explicitly identifies the model's development, capabilities, timing, availability and expected benefits as forward-looking statements that may differ from actual results.

Sources & further reading

• OpenAI and Synopsys announce GPT-Synopsys. Synopsys, September 30, 2026. The primary announcement documents the agreement, proposed workflow, hosted service and early engagements. Its forward-looking statement covers development, timing, availability, capabilities and benefits. It gives no public GPT-Synopsys launch date, price or measured customer result.
• Synopsys announces AgentEngineer and Autopilot. Synopsys, September 28, 2026. This separate release says more than 50 engagements are underway for that portfolio and places its planned availability at the end of 2026. Its results and timetable should not be assigned to GPT-Synopsys.
• Synopsys.ai Copilot product overview. Synopsys, accessed September 30, 2026. The company describes five commercially available assistants and advisors and its own early feedback on them. This is context for existing products, not independent validation of the new model.