OpenAI and Synopsys Sign Chip-Design Model Deal With No Customers or Benchmarks Named
A.I. / news
OpenAI and Synopsys Sign Chip-Design Model Deal With No Customers or Benchmarks Named
GPT-Synopsys will run Synopsys EDA tools on OpenAI-hosted infrastructure under a revenue-sharing agreement. The Sept. 30 announcement gives no dollar figure and no ship date.

OpenAI and Synopsys announced on Sept. 30 a multi-year agreement to build GPT-Synopsys, a model that operates chip-design software directly, with revenue split between the two companies and no dollar figure disclosed.
The announcement came from Synopsys, which makes electronic design automation (EDA) software, the programs engineers use to lay out and verify chips. Early technology engagements with semiconductor customers are already underway, Synopsys said in its press release.
What Synopsys and OpenAI announced on Sept. 30
GPT-Synopsys pairs OpenAI models with Synopsys EDA tools. Engineers hand it a design objective, and the model runs the tools, reads their output, makes changes and iterates until a result is ready for human review.
OpenAI will license Synopsys' EDA tools as part of the deal. The model will run on OpenAI-hosted infrastructure, and Synopsys said customer design data is not used for model training. Data is encrypted at rest and in transit.
Sassine Ghazi, president and CEO of Synopsys, said the industry needs faster design without giving up power, performance and area targets or first-time-right silicon. Greg Brockman, president and co-founder of OpenAI, said: "We're using our most advanced technology to improve the systems that power AI."
What the Sept. 30 announcement leaves out
The release contains no benchmark, no named customer, no price and no availability date for anyone beyond the early engagements.
| Item | In the announcement? |
|---|---|
| Multi-year term | Yes, length not given |
| Revenue sharing | Yes, split not given |
| OpenAI licensing Synopsys EDA tools | Yes |
| Customer names | No |
| Performance or timing results | No |
| General availability date | No |
The revenue-sharing line is the only financial term. Neither company said which side books the revenue from a customer's subscription, or whether OpenAI pays for the tool licences up front.
Synopsys gave no result for any real chip. Until it does, claims that the model shortens a design cycle rest on the companies' description of intent rather than on measured outcomes.
Why verification stays with conventional tools
Semiwiki, a site covering the EDA industry, wrote that "conventional engineering tools will continue checking the model's work against physical requirements." That places the model upstream of sign-off checks that already exist, rather than replacing them.
Semiwiki also described Synopsys as the "top EDA company AND the top IP company", and wrote that engineering roles may shift toward architecture, requirements and supervising AI.
Synopsys describes the same arrangement: agents work toward results that engineers then review.
The deal lands two days after OpenAI published safety-case guidelines for frontier training on Sept. 28. Separately, chip startup Efficient Computer raised over $97 million at a $650 million valuation on a data-center energy claim that The Terminal reported as unproven.
Neither company said when GPT-Synopsys will be generally available. Synopsys said early engagements are underway, and the announcement does not say when the first results will be published.
Sources
More in A.I.
- 01Runta's 27,280-Run Harness Eval Finds the Model Explains 92.8% of Agent CostFrontierHarness v1.1 ran 22 coding-agent harnesses against five open models, and Codex CLI went from first place to sixteenth when the model changed.
- 02Qwen3.8-27B Has 6.9 Million Downloads on Hugging Face. Artificial Analysis Rates It Slow and VerboseAlibaba's Apache 2.0 dense model beats its size class on an independent index of 20, and takes 38.8 minutes a task against a 12.6-minute median.
- 03Claude Code Team Publishes Its Verification Rules: 39 Checks, a Three-Round Limit and Observe ModeA Nov. 16 post says about 90% of the team's code is written and reviewed by Claude, and lays out the gates that keep it shippable.
- 04Inkling 2 Small Weights Land on Hugging Face Under Apache 2.0, and the Card Asks for 500GB of VRAMThinking Machines opened the weights a week after launching on Tinker. The safety section still says release is pending.