TL;DR: Synopsys and OpenAI signed a multi-year agreement on Sept. 30, 2026 to jointly develop GPT-Synopsys, a specialized model tuned to operate Synopsys electronic design automation tools for semiconductor design workflows under a shared revenue framework.

Partnership announced in San Francisco and Sunnyvale

Synopsys (Nasdaq: SNPS) and OpenAI announced a strategic partnership on Sept. 30, 2026, according to the Synopsys announcement titled "OpenAI and Synopsys Announce GPT-Synopsys: Frontier Intelligence to Revolutionize Chip Design."

The announcement was issued from San Francisco and Sunnyvale, Calif., on Sept. 30, 2026, and described the companies as preferred partners for development and delivery of the specialized model.

Reuters reported the same agreement on Sept. 30, 2026, in the article titled "Synopsys, OpenAI strike deal to develop AI model for chip design work," with reporting by Stephen Nellis and a timestamp of 7:01 PM UTC.

Reuters described Synopsys as a maker of software used in designing computing chips and said the company disclosed the agreement on Wednesday.

The two reports align on the central transaction: a joint development effort for a chip-design model called GPT-Synopsys, paired with shared revenue terms.

What GPT-Synopsys is intended to do

GPT-Synopsys is described as a specialized model for chip design that combines OpenAI frontier AI with Synopsys electronic design automation technology and chip-design expertise.

The Synopsys announcement said the specialized model is optimized to use Synopsys EDA tools to perform semiconductor design workflows.

Reuters said GPT-Synopsys will be specifically trained and tuned to take on chip design tasks.

Those tasks were described by Reuters as beginning with a description of a chip's circuits in a code-like language and extending to determining how to lay out billions of transistors on a small square of silicon.

The Synopsys announcement said GPT-Synopsys brings together OpenAI frontier models with Synopsys EDA technology and domain expertise.

That combination is intended to enable the specialized model to reason about chip design and verification and to directly operate Synopsys tools.

From connected tools to expert operation

The Synopsys announcement distinguished current practice from the planned advance.

It said agentic AI technologies currently connect general-purpose models to EDA tools to run chip-design workflows.

The next step described in the announcement is to make frontier models specialists in using EDA tools.

That specialization involves learning to run the tools as expert engineers would, interpreting tool outputs, and iteratively optimizing designs through the tools.

The announcement said this integration will make the model a native expert user of the tools.

The stated engineering purpose is to help engineers consider more design alternatives and improve power, performance and area, a metric abbreviated as PPA, in order to deliver more sophisticated designs more rapidly.

How engineers would delegate work

Engineers would assign design objectives to agents built around GPT-Synopsys, according to the Synopsys announcement.

Delegated objectives would include PPA optimization and work extending to timing and verification closure.

Agents would run tools, interpret results, implement changes, and iterate toward verified outcomes for engineer review.

The announcement framed that loop as tool execution followed by result interpretation, design revision, and further iteration until an outcome is ready for review by an engineer.

The announcement also said the model is being developed to reason about both chip design and verification.

Direct operation of Synopsys tools by the model is a central element of that workflow description.

Multi-year business and delivery structure

The companies signed an expansive, multi-year agreement to jointly develop GPT-Synopsys, according to the Synopsys announcement.

The agreement includes a revenue-sharing arrangement.

It also includes go-to-market collaboration to make GPT-Synopsys available to customers worldwide.

Reuters likewise reported that Synopsys said it struck a deal to share revenue with OpenAI as the two develop an AI model for the chip business.

The Synopsys announcement said the companies will closely collaborate on research and development and on joint go-to-market initiatives with a shared revenue framework.

The joint service offering described in the announcement would provide bundled compute, model access, and licenses.

OpenAI licensing of Synopsys tools

OpenAI will license Synopsys electronic design automation tools for development of the specialized model, according to the Synopsys announcement.

That licensing element was listed among the principal highlights of the transaction.

The licensed tools provide the development environment in which the model would learn tool operation for semiconductor design workflows.

The announcement linked that access to development of GPT-Synopsys rather than to general model training.

It presented licensed EDA tools, frontier AI capabilities, and chip-design expertise as the three inputs brought together for the project.

Hosting, integration and customer systems

GPT-Synopsys will run on OpenAI-hosted infrastructure, according to the Synopsys announcement.

It will be designed to interoperate with customer agent harness systems.

It will also be deeply integrated with Synopsys.ai and with Synopsys Autopilot, described as an agentic AI platform.

The joint offering is intended to combine compute, model, and license elements in one service structure.

The announcement said early technology engagements are underway with leading semiconductor customers.

It did not name those customers in the published announcement text.

Protection of customer design data

The joint service offering is structured to protect customer-specific design data while supplying bundled compute, model, and licenses, according to the Synopsys announcement.

GPT-Synopsys will feature enterprise-grade security, governance, and access controls.

Customer data is not used to train the model, according to the announcement.

Customer data is encrypted at rest and in transit.

Retention, audit, and permission controls can be configured by customers for management of that data.

The announcement presented those controls as part of the enterprise framework for the joint offering.

Leadership statements on manufacturing rigor

Sassine Ghazi, president and chief executive of Synopsys, said semiconductor engineering needs faster chip-design work without compromising PPA or first-time-right silicon, according to the Synopsys announcement.

Greg Brockman, president and co-founder of OpenAI, said the companies are applying advanced technology to chip design to help engineers review more designs and reach a working chip faster, according to the same announcement.

The Synopsys announcement quoted Ghazi saying the agreement will expand access to Synopsys advanced design capabilities and engineering tools required for increasingly complex chips.

Brockman linked improved chips to improved AI systems and broader availability of AI.

What the companies labeled as forward-looking

The Synopsys announcement included forward-looking statements covering development, capabilities, timing, availability, and anticipated benefits of GPT-Synopsys and the strategic partnership.

It said those statements involve risks, uncertainties, and other factors that could cause actual results, time frames, or achievements to differ materially from those expressed or implied.

It directed readers to risk information in filings Synopsys makes with the SEC, including sections entitled "Risk Factors" in its latest Annual Report on Form 10-K and latest Quarterly Report on Form 10-Q.

It said Synopsys undertakes no duty and does not intend to update any forward-looking statement in the announcement unless required by law, whether because of new information, future events, or otherwise.

That caution means capabilities, timing, availability, and benefits described for GPT-Synopsys remain planned outcomes rather than completed results in the Sept. 30, 2026 disclosures.

Sources referenced