Between Silicon and Intelligence: Synopsys and OpenAI Bring Artificial Intelligence Into the Architecture of Future Chips

A modern semiconductor begins as an idea expressed through layers of code, mathematics and engineering decisions. Long before a finished chip enters a computer, vehicle or data center, thousands of choices must be made about circuits and the arrangement of billions of transistors.

Synopsys and OpenAI are now attempting to bring artificial intelligence deeper into that process. The companies announced an agreement to develop a specialized model called GPT-Synopsys for semiconductor design work.

The model is intended to learn how to use Synopsys software tools and assist engineers with the complex sequence of decisions involved in chip development. Those tasks range from describing circuits to determining how enormous numbers of transistors should be positioned on a small piece of silicon.

OpenAI co-founder Greg Brockman said the system could help engineers move through design trade-offs and optimization processes more quickly, potentially reducing the time required to complete certain stages of development.

The partnership comes as semiconductor design becomes increasingly important to the expansion of artificial intelligence itself. Advanced AI systems require powerful processors, while those processors depend on increasingly sophisticated design software and engineering methods.

That creates a circular relationship between AI and semiconductor technology. Artificial intelligence requires better chips, while better chips can increasingly be designed with the assistance of artificial intelligence.

Synopsys said the arrangement includes a training subscription fee paid by OpenAI. Once customers begin using the resulting product, the two companies will share revenue according to the model’s effectiveness in improving chip design.

The companies also emphasized that AI-generated design work will not simply move directly into production. Synopsys tools using traditional computing techniques will continue to verify whether designs function correctly and satisfy physical requirements. This verification process is often referred to as “sign-off.”

That additional layer is important because semiconductor manufacturing allows very little room for error. A design that appears plausible to an AI system still needs to be tested against physical rules and engineering requirements before it can become a functioning chip.

The announcement came as Synopsys also raised its fiscal 2027 revenue-growth forecast to 15%, above the 11.19% analyst expectation cited by Reuters. Its shares rose as much as 7% following the partnership and forecast.

The development points toward a broader shift in the semiconductor industry. AI is no longer being discussed only as the software running on chips; it is increasingly becoming part of the process used to design the chips themselves. The result could be a technology cycle in which intelligence and silicon become progressively more intertwined.