# OpenAI and Synopsys Announce GPT-Synopsys, a Model That Runs Chip-Design Software. What's Confirmed, and What Isn't Priced or Dated Yet

The two companies will build a model trained to operate Synopsys design tools like an expert user, sold as a joint service with revenue sharing. No price and no general availability date have been given.

By Priya Nair (AI Industry Reporter) — published 2026-10-03, updated 2026-10-03
Source: https://aiscoutdaily.com/news/openai-synopsys-gpt-synopsys-chip-design-model

## TL;DR
OpenAI and Synopsys announced a multi-year partnership on September 30, 2026 to build GPT-Synopsys, a model that operates electronic design automation (EDA) software to run semiconductor design and verification workflows. It will be delivered as a joint service on OpenAI-hosted infrastructure, with joint go-to-market and revenue sharing. Customer design data will not be used to train the model. Early engagements with chip customers are under way, but the announcement gives no price, no general availability date, and no performance figures.

Chip design runs on specialist software, and most of the work is engineers driving that software, reading what it prints, and trying again. OpenAI and Synopsys, the maker of widely used EDA tools, say they will train a model to do that driving.

## What was announced

- A multi-year strategic agreement to develop GPT-Synopsys, a specialised model that operates Synopsys EDA tools directly.
- The model will be trained to act as an expert user: run design workflows, interpret tool output, make changes, and iterate toward goals such as power, performance and area, timing closure, and verification.
- OpenAI will license Synopsys EDA tools to develop the model. The companies will run joint research, go-to-market, and revenue sharing; the announcement does not give the split.
- The service will run on OpenAI-hosted infrastructure and combine compute, the model, and Synopsys software licences. It is meant to work with Synopsys.ai and Synopsys Autopilot and with customers' own agent harnesses.
- Engineers set objectives and review results, while agents carry out part of the workflow.

## Data and status

Per the announcement, customer design data will not be used to train the model. The service is described as having encryption at rest and in transit, plus configurable retention and permission controls. Early technology engagements with semiconductor customers are already under way, but no general availability date is given.

## What's missing

| Detail | Status |
| --- | --- |
| Price or licensing model for customers | Not disclosed |
| Revenue-sharing split | Not disclosed |
| Benchmarks or task success rates | None given |
| General availability | No date; early customer engagements only |
| Which OpenAI model it is built on | Not stated in the material we read |

*From the September 30 announcement.*

AI Weekly's roundup reports that Synopsys shares closed up on announcement day; that reflects investor interest, not product evidence.

## What this means for tool buyers

This is a vertical agent: a general model paired with one vendor's licensed tools and sold as a bundle. If you buy EDA software, expect your Synopsys rep to raise it, and ask for a measured task, such as closing timing on one of your own blocks, before committing budget. For everyone else, it is another sign that the agent products getting money behind them are the ones tied to a specific professional toolchain rather than a general chat window.

## Sources
- [Synopsys and OpenAI announcement, via StockTitan (Sep 30, 2026)](https://www.stocktitan.net/news/SNPS/open-ai-and-synopsys-announce-gpt-synopsys-frontier-intelligence-to-0perv30u66fs.html)
- [AI Weekly: AI News Today, October 2](https://aiweekly.co/ai-news-today)