OpenAI and Synopsys have signed a multi-year strategic partnership to develop GPT-Synopsys, a specialised artificial intelligence model designed to assist with computer chip architecture. Synopsys supplies the electronic design automation tools required for creating semiconductors, while OpenAI provides the underlying machine learning technology. The new system will run on OpenAI infrastructure and will not use customer data for training, instead storing all information in an encrypted format. Early trials with semiconductor clients are currently in progress, with both firms agreeing to market the solution jointly and share revenue from its use.
The collaboration addresses the growing need for faster chip development cycles without compromising engineering oversight. While AI models can generate code, human engineers remain responsible for reviewing and approving final outputs to ensure reliability. This approach aims to reduce manual workload while maintaining strict control over complex hardware specifications. The project builds on existing work between OpenAI and Broadcom, which recently launched the Jalapeno chip for artificial intelligence workloads.
- All customer data remains encrypted and excluded from model training
- Human engineers retain full authority to review and approve designs
- Revenue sharing and marketing responsibilities are split between both companies
