THE ESSENTIALS IN 30 SECONDS

  • Robotics researchers now report to Chief Research Officer Mark Chen; applied robotics and hardware staff moved to President Greg Brockman's Applied organisation.
  • OpenAI confirmed the change to Fortune and said robotics remains core to its long-term vision, describing the move as an acceleration, not a retreat.
  • Fortune cites an Nvidia study in which a GPT-6 Astra-powered agent scored 74.8% on a simulated manipulation benchmark, against 51.0% for the best specialised robot models.

OpenAI no longer has a separate robotics team. In an exclusive published on 13 November, Fortune reported that the company has dissolved the unit and reassigned its members to other parts of the organisation. OpenAI confirmed the reorganisation when asked by the publication.

Where the staff went

According to Fortune, the robotics research group has joined OpenAI's main research team under Chief Research Officer Mark Chen. The applied robotics and hardware teams have moved into the broader Applied organisation led by President Greg Brockman. Fortune did not say how many people were affected.

An OpenAI spokesperson told the publication that robotics remains central to the company's long-term vision and that the change is meant to speed up progress rather than signal a pullback. Fortune also noted that a booth OpenAI had registered at CoRL, the academic robotics conference held in Austin, stood empty when its reporter visited.

The bet on general models

The rationale, as described in the report, is a shift in how OpenAI expects robotics to advance. Earlier this year, researchers thought progress would come from building on video models such as Sora, which learn something about how the physical world behaves. Since then, OpenAI believes its general frontier models, including GPT-6 Astra, can be applied to robots directly.

Fortune points to a study by Nvidia in which an agent running on Astra reached a 74.8% success rate on a simulated manipulation benchmark, while the best specialised robotics models on the leaderboard reported 51.0%. These are simulation results; the article does not describe real-world tests. OpenAI is still developing its own humanoid robot, and Fortune reports that CEO Sam Altman is aiming for a first public demonstration in 2027.

Context and caveats

This is not OpenAI's first change of course on robotics. It wound down an earlier team in 2020, rebuilt one in 2024 alongside a partnership with Figure, and lost that partner in February 2025 when Figure chose to build its own robot AI. Fortune links the latest move to cost pressure, citing $10.9 billion in operating losses last quarter and reports that Altman asked staff to drop "side quests".

Researchers quoted by Fortune were cautious. Carnegie Mellon's Guanya Shi said smaller models running on the robot itself will still be needed, since large cloud-hosted models are slow and depend on a constant connection. Details of the reorganisation come from Fortune's reporting and OpenAI's statement to it; the company has not published its own announcement.