OpenAI and Synopsys plan a model for chip design
The partners plan GPT Synopsys for chip design with early customer work and no announced public release date
OpenAI and Synopsys announced a September 30 agreement to develop GPT Synopsys, a specialized model intended to operate Synopsys electronic design automation tools. Early customer engagements are underway.
Tools would remain at the center
Agents would run design tools, interpret results and revise designs toward engineering targets such as power, performance and area. Engineers would review the outcomes.
The proposed service would run on OpenAI infrastructure and bundle compute, the model and software licenses. Synopsys.ai and Autopilot integration is planned.
Availability remains an open question
The announcement gives no public release date or pricing and supplies no measured results for GPT Synopsys.
Synopsys says customer design data will not train the model, with encryption and configurable retention, audit and permission controls planned.
What an engineering trial should establish
A useful evaluation would start with a fixed design and agreed acceptance criteria. Compare the final design, compute consumed and engineering review time against the existing workflow. Keep a record of every constraint and design change the agent makes. A quicker intermediate result only helps if the final checks still pass.
This reporting does not include a product test.
Illustrative silicon wafer photograph by Laura Ockel, published in April 2020, under the Unsplash License. This image does not show a GPT Synopsys result. Source supplied JPEG rendition with no local edits. No endorsement is implied.
