OpenAI has broken up its standalone robotics team. The company confirmed the reorganisation to Fortune on 13 November 2026: its robotics researchers now sit inside the main research group, and its applied robotics and hardware staff have moved into the product side. OpenAI insists this is not a retreat from robotics. It is a bet — that its general-purpose AI models can now run robots without a dedicated team behind them.
Where the robotics staff have gone
Under the reorganisation, the robotics research team reports to Chief Research Officer Mark Chen, inside OpenAI's central research organisation. The applied robotics and hardware teams have joined the larger Applied organisation, which reports to President Greg Brockman. OpenAI has not published a statement or blog post of its own about the change, and it has not said how many people were affected. The move came so recently that OpenAI had already booked a stand at CoRL, the leading academic robotics conference in Austin, Texas, this week — and the booth sat empty.
Why OpenAI thinks it can do this
The bet rests on GPT-6 Astra, OpenAI's general-purpose model. According to OpenAI, its researchers concluded at the start of 2026 that the best route to capable robots ran through models like Sora, its video generator — which has to grasp something of how the physical world works to make realistic clips — and now through Astra itself. In a research paper, "An Unexpected Robot Policy: Early Evaluations of GPT-6 Astra on RoboDojo and Beyond", an Astra-based agent reached a 74.8% success rate on an Nvidia robotics benchmark, against 51% for leading specialised models. OpenAI's argument, in short: the model is now good enough at understanding the physical world that a separate robotics organisation is no longer the fastest way forward.
A team that has been built, shut down and rebuilt before
This is the second time OpenAI has broken up a robotics team. It had one in its early years, then wound it down in 2020 — co-founder Wojciech Zaremba later cited limited robot data and a more efficient path toward artificial general intelligence. It rebuilt a standalone team in 2024, announcing a partnership with the humanoid-robot startup Figure to develop AI models for humanoid robots. That partnership ended in February 2025, when Figure said it would develop its own robot AI. OpenAI has said it is currently developing its own humanoid robot, with CEO Sam Altman targeting a first public demonstration in 2027 — a target the company has not changed.
What it means
- For OpenAI: robotics becomes a capability of the frontier model, not a separate division. If Astra keeps improving, the reorganisation looks prescient; if robot control turns out to need robotics-specific work after all, the company has just scattered the people who do it.
- For the robotics field: the biggest AI lab in the world is now betting that general models beat specialised robot stacks. Every robotics startup that sells a specialised control system just got its thesis challenged by the company with the most to lose.
- For Figure and the humanoid race: OpenAI's former partner is now a direct rival building its own robot AI, while OpenAI still aims at a 2027 humanoid demo — now without a standalone team behind it.
Bottom line
OpenAI has not quit robotics. It has dissolved the team and handed the job to its most powerful model. The 74.8% benchmark number is the evidence it cites; the empty booth at CoRL is the awkward footnote. The real test comes in 2027, when Altman has promised a humanoid robot on stage — built, for the first time, by a company with no robotics team at all.
Sources: Fortune's exclusive report (13 November 2026) as carried by multiple outlets; OpenAI's GPT-6 Astra robotics paper. Hero photo: a humanoid robot being programmed, by ZhanatKappassov, CC0 public domain, via Wikimedia Commons.
