MLQ.ai (Fortune) | 2026-11-13

OpenAI has disbanded its standalone robotics team and moved its members onto other teams, Fortune reported on Friday, Nov. 13 [1]. The robotics research group now sits inside OpenAI's main research team, reporting to chief research officer Mark Chen, while the applied robotics and hardware teams have moved to the Applied organization, which reports to president Greg Brockman [1]. Fortune did not report how many people were affected.

OpenAI tied the change to how its research has gone. Its researchers first assumed that building on video models such as Sora was the best route into robotics, but its general-purpose frontier models have since shown they can be applied to robots directly, the company told Fortune [1]. A spokesperson said robotics "remains a core part of OpenAI's long-term vision," according to Fortune's account [1]. OpenAI is developing its own humanoid robot, and Sam Altman is targeting a first public demonstration in 2027 [1].

In the experiments Fortune described, researchers gave GPT-6 Astra camera images, information about the robot and a control interface for commanding its actions [1]. In a study run by Nvidia, an Astra-powered agent reached a 74.8% success rate on a simulated manipulation benchmark, where the model commands a simulated robot, compared with 51.0% for the best specialized robotics models on that leaderboard [1].

Robotics researchers quoted by Fortune were split on what that means. Guanya Shi, an assistant professor at Carnegie Mellon's Robotics Institute, said the move suggests robotics is "no longer a special problem" but a downstream application of general intelligence, though he still sees a need for small onboard models because cloud-hosted language models are slow [1]. Russ Tedrake, the MIT professor who co-founded Walden Robotics, argued that robot makers keep an edge from real-world deployment data, and Pete Florence, co-founder of robot-model startup Generalist, said it is unclear any existing model will dominate [1]. Google DeepMind took the specialized route this week, releasing its Gemini Robotics 2.5 vision-language-action model to partners (our story).

This is the second time OpenAI has broken up a robotics group. It wound down its first team in 2020, with co-founder Wojciech Zaremba later citing a shortage of robot data, and rebuilt a standalone team in 2024 alongside a partnership with Figure [1]. Figure ended that partnership in February 2025 to develop its own robot AI [1]. Former members of the team were at OpenAI's booth at the CoRL robot learning conference in Austin this week, without a demo [1]. The reorganization comes as OpenAI faces cost pressure ahead of a possible IPO in early 2027, with reported operating losses of $10.9 billion last quarter [1].

Read more