OpenAI moves ChatGPT and Codex into physical engineering
The company demonstrated AI-assisted CAD, circuit-board and simulation workflows with internal teams, Mazda and Chip Ganassi Racing.
OpenAI is expanding ChatGPT and Codex from software into mechanical and electrical engineering, with early workflows for CAD, circuit boards, simulation and manufacturing documentation.
In one demonstration, two OpenAI employees used GPT-6 Astra to define requirements and design the enclosure, printed circuit board, interface, firmware and software for a device that talks to Codex. They completed the design in under a week as a side project, and a contract manufacturer produced a small batch of prototypes. Human review was still necessary because the model missed production limits, tolerances and assembly details.
OpenAI says its hardware teams already use ChatGPT and Codex to revise CAD models, prepare drawings, set up simulations and review designs. Model usage across those teams has grown 23-fold since February.
The company is also working with Chip Ganassi Racing to turn 3D scans of existing parts into editable CAD and manufacturing drawings. A separate research collaboration with Mazda focuses on airflow and structural simulations for new vehicle designs.
This is not a standalone engineering product launch. It is a capability update and a call for engineering firms and software vendors to collaborate. The demos nonetheless show frontier models moving from code generation into tightly coupled hardware, software and manufacturing workflows.
OpenAI
