OpenAI disbands its robotics team as its general models start running robots

OpenAI has disbanded its stand-alone robotics team, moving its research staff into the company’s main research group and its applied robotics and hardware staff into the broader Applied organization. The company says robotics remains part of its long-term plans and argues the reshuffle will speed development of general-purpose robots. The change follows a shift in strategy: OpenAI had once pursued robotics separately, then rebuilt a dedicated team in 2024 and partnered with humanoid-maker Figure before that partnership ended in 2025. CEO Sam Altman has said OpenAI is developing its own humanoid robot, with a public demonstration targeted for 2027. The company now believes general models such as GPT-6 Astra can be applied directly to robots, rather than requiring a separate robotics-focused AI system.

The shift reflects a wider bet that language, vision, and reasoning models can control machines in the physical world. In one simulated manipulation test, an Astra-powered agent scored 74.8% success, compared with 51% for the leading specialized model cited. Researchers say general models can adapt to changes in object position or lighting and reason through task steps, but they can be slow, cloud-dependent, and reliant on a network connection. Robotics specialists also retain an advantage: robots operating in factories collect practical, task-specific experience and data. The move may challenge startups built around dedicated robot “brains,” while also aligning with OpenAI’s pressure to reduce costs and focus on core products amid heavy losses and plans for a possible IPO.