On 13 November 2026, Fortune reported that OpenAI has disbanded its stand-alone robotics team and moved the researchers to other groups. An OpenAI spokesperson confirmed the change to Fortune and said that “robotics remains a core part of OpenAI’s long-term vision.” For robotics labs, university partners and research offices, the practical question is whether general-purpose models are replacing dedicated robot-learning models as the default research path.
What was reported
According to Fortune’s exclusive by Wen Shao, published on 13 November 2026:
- The robotics research team moved into OpenAI’s main research organization, which reports to Chief Research Officer Mark Chen.
- The applied robotics and hardware teams moved into the Applied organization, which reports to President Greg Brockman.
- OpenAI says the reorganization is meant to speed progress on general-purpose robots, not to end robotics work.
OpenAI’s stated rationale
Fortune reports that OpenAI’s researchers first assumed that building on video models such as Sora was the best route to robotics, and later found that general frontier models, including GPT-6 Astra, can be applied directly to robot control. Fortune also cites an Nvidia study in which an Astra-powered agent reached a 74.8% success rate on a simulated manipulation benchmark, against 51.0% for the best specialized robotics models on that leaderboard. That figure is reported by Fortune from the Nvidia study and is a simulated benchmark result, not evidence about performance on physical hardware in a lab.
History and hardware plans
Fortune notes that OpenAI wound down an earlier robotics team in 2020, when cofounder Wojciech Zaremba cited limited robot data. It also reports that a 2024 partnership with the startup Figure, to build AI models for humanoid robots, ended in February 2025 when Figure said it would develop its own robot AI. Fortune adds that OpenAI is developing its own humanoid robot, with CEO Sam Altman targeting a first public demonstration in 2027.
What outside researchers said
Fortune quotes several academics and practitioners with differing views:
- Guanya Shi of Carnegie Mellon University said robotics may now be treated as “no longer a special problem,” a downstream application of general intelligence. He also still sees a need for smaller models that run onboard without constant connectivity.
- Russ Tedrake of MIT and Walden Robotics argued that robotics startups keep an edge through real-world deployment data.
- Pete Florence of Generalist said it is unclear which model will dominate robotics in the long term.
What it means for research labs and research offices
The following is editorial interpretation rather than reporting, and the picture is still developing.
- Benchmarks and methods. If general models keep outperforming specialized ones on shared benchmarks, labs working on robot learning may need to justify specialized-model approaches against general-model baselines in proposals and papers.
- Partnerships. Industry collaborations built around a named company team can change when that team is reorganized. Research offices may want agreements to specify deliverables, data and points of contact so that they survive a restructuring. The Figure partnership’s 2025 end, as reported by Fortune, is one example of how such arrangements shift.
- Deployment data. The view that real-world data remains a moat points to the value of physical lab setups and data-sharing terms, including in lab automation.
- Evidence quality. The headline benchmark figure is a simulated result relayed through press coverage, so claims built on it deserve a check against the original Nvidia study before they enter a grant narrative.
What is not yet known
The reporting does not say how many researchers were affected, whether existing external research collaborations change, or how OpenAI will release or document its robotics work. OpenAI’s own detailed statement beyond the spokesperson’s comment was not part of the Fortune account.
Sources
- Fortune, “Exclusive: OpenAI disbands its stand-alone robotics team as its general-purpose AI models start running robots,” 13 November 2026, https://fortune.com/2026/11/13/openai-disbands-robotics-team-as-its-general-models-start-running-robots/. Read 2026-11-14.
