OpenAI and chip-design software company Synopsys have signed a multi-year agreement to develop GPT-Synopsys, a specialized AI model designed to use Synopsys’ electronic design automation tools. The companies say it will help engineers run chip-design workflows, review tool results and adjust designs through repeated iterations.
The planned system is intended to work on tasks such as optimizing power, performance and area, and completing timing and verification work. Engineers would set objectives and review the results; AI agents would run the tools, interpret their output and make proposed changes. The announcement describes a product in development, not a service with a stated launch date.
How the service is planned to work
GPT-Synopsys is expected to run on OpenAI-hosted infrastructure and connect with customer agent systems. Synopsys also plans to integrate it with Synopsys.ai and its Autopilot agent platform. The companies say the offering will bundle computing capacity, the model and software licenses.
The companies say customer-specific design data will not be used to train the model. They also say data will be encrypted while stored and transferred, with configurable controls for retention, auditing and permissions. These are announced plans; the release does not provide independent security details or results from customer use.
What is known so far
Synopsys and OpenAI say early technology engagements with leading semiconductor customers are already underway. They also plan joint research and development, sales efforts and a shared revenue arrangement. The announcement does not name the customers, give pricing, or say when other companies can access the service.
Why it matters
For chip-design engineers, the proposed tool could take on repeated software runs and design adjustments, while leaving engineers to set goals and review outcomes. The companies have not yet shared access dates, prices or evidence of how well it performs in customer projects, so its practical value is still unknown.
