OpenAI and Synopsys have entered a partnership to develop GPT-Synopsys, an AI model designed for autonomous chip design. The system will integrate OpenAI's frontier models with Synopsys’s EDA software, allowing the AI to directly operate engineering tools, interpret results, and perform iterative design tasks like PPA optimization.
Under the agreement, OpenAI will license Synopsys tools to build the model on its own infrastructure. The companies plan to jointly release the product and share the resulting revenue. While the model will be trained using Synopsys domain expertise, the companies stated that customer design data will be excluded from future model training.
Synopsys has been expanding its AI portfolio, including the launch of DSO.ai in March 2020 and Synopsys.ai Copilot in November 2023. In September 2024, the company introduced its Autopilot platform and a suite of seven agents. Early technology engagements with semiconductor customers are currently underway.
The collaboration follows OpenAI's internal development of the Jalapeño inference accelerator, created in partnership with Broadcom. By utilizing its own AI models, OpenAI completed the design-to-tape-out process for that chip in nine months. The new GPT-Synopsys product aims to scale these capabilities for professional design teams.
The companies have not yet disclosed a specific release date or pricing structure for the service. The project faces competition from industry rivals, such as Cadence, which launched its ChipStack AI Super Agent in February and announced plans for Level-5 autonomy integration by the second half of 2026.
