Particle.news

OpenAI and Synopsys Form Partnership to Build GPT‑Synopsys for Chip Design

The agreement places OpenAI models on OpenAI‑hosted compute integrated with Synopsys EDA tools under enterprise security to speed semiconductor design.

Overview

  • The companies announced the multi‑year collaboration on Sept. 30, 2026, confirming they will jointly develop a specialized model called GPT‑Synopsys that is tuned to run electronic design automation (EDA) workflows.
  • Under the deal OpenAI will license Synopsys’ EDA tools for training and development while Synopsys and OpenAI will share revenue when customers use the joint product.
  • GPT‑Synopsys will run on OpenAI‑hosted infrastructure, bundle compute, model access and Synopsys licenses for customers, and include controls that encrypt data and forbid using customer design data to train the model.
  • Both companies say early technology engagements are already underway with leading chip customers and that the model will act as an agent that runs tools, reads outputs, and iterates on designs subject to verification by Synopsys’ traditional sign‑off tools.
  • The announcement builds on Synopsys’ prior AI moves such as Synopsys.ai Copilot on Azure and its autonomous workflows with AMD and Microsoft and follows major industry investments including Nvidia’s $2 billion commitment to Synopsys.