OpenAI disbands its stand-alone robotics team and moves the researchers into its main groups

OpenAI has disbanded its stand-alone robotics team, Fortune reported on 13 November 2026. The robotics researchers moved into the main research team, which reports to Chief Research Officer Mark Chen. The applied robotics and hardware staff moved into the Applied organisation under President Greg Brockman. An OpenAI spokesperson confirmed the change to Fortune and said robotics “remains a core part of OpenAI’s long-term vision”. Fortune gives no figure for the team’s size and does not name a leader.

The reason given is that OpenAI’s general frontier models, GPT-6 Astra among them, have turned out to apply directly to robotics. Earlier in 2026, Fortune says, researchers expected the best route to be building on video models such as Sora, which had some grasp of how the physical world behaves.

OpenAI wound down its first robotics team in 2020, citing a shortage of robot data, and rebuilt one in 2024 alongside a partnership with Figure that ended in February 2025. Fortune says OpenAI is developing its own humanoid robot, with a first public demonstration planned for 2027.

Researchers quoted by Fortune are split. Guanya Shi of Carnegie Mellon says robots will still need small models that run on board, since large language models are slow and need the cloud. Russ Tedrake of MIT says robotics companies keep an edge through the data their deployed robots collect.

What it means

OpenAI is betting that the same general model that writes and codes can also drive robots, so it no longer runs a separate robotics lab. Robotics startups that build their own robot AI are betting the other way.

Caveats

OpenAI has published no results showing GPT-6 Astra controlling a robot. The claim that general models apply directly to robotics comes from the company’s account to Fortune.
The confirmation is a spokesperson quoted by Fortune. OpenAI has not posted about the change.
Fortune also links the move to OpenAI’s work to cut costs and drop side projects, so part of it may be a budget decision.
Fortune cites an Nvidia study of Astra on a simulated robot benchmark but names neither the study nor the benchmark, so its scores are not repeated here.