OpenAI disbands its robotics team as GPT-6 Astra starts running robots

AI NewsIshaan VardhanNov 14, 20265 min read

OpenAI has disbanded its stand-alone robotics team and moved its staff into the company’s main research and applied divisions, Fortune reported on Friday, 13 November 2026, in a move OpenAI confirmed and said would speed up robotics work rather than end it. The reason matters well beyond San Francisco: OpenAI now believes general-purpose models such as GPT-6 Astra can run robots directly, which changes what robotics startups in India and elsewhere need to build themselves.

Key takeaways

• OpenAI’s robotics researchers now sit in the main research team under Chief Research Officer Mark Chen; applied robotics and hardware staff moved to the Applied organisation under President Greg Brockman.
• OpenAI says robotics stays part of its long-term plan, and CEO Sam Altman is still targeting a first public demonstration of its own humanoid robot in 2027.
• The bet is that one general model can control robots: in an Nvidia study cited by Fortune, an agent built on GPT-6 Astra reached a 74.8% success rate on a simulated manipulation benchmark.
• This is the second time OpenAI has shut a dedicated robotics team; the first closed in 2020.

What OpenAI changed

OpenAI no longer has a separate robotics team. Its robotics research staff were folded into the core research group that reports to Mark Chen, and the people working on applied robotics and hardware moved into the Applied organisation that reports to Greg Brockman, according to Fortune’s report by Wen Shao. Fortune did not give a headcount. Fortune also noted that OpenAI had registered a booth at CoRL, an academic robotics conference in Austin, and that the booth was empty when its reporter visited.

The story spread quickly after Fortune published it: Techmeme put it on its front page, and AI newsletters and finance sites, including Seeking Alpha, picked it up within hours.

Why a general model is replacing a robotics team

OpenAI’s reasoning is that its general-purpose models are now good enough to drive robots without a separate robotics-specific model. Fortune reported that OpenAI researchers found GPT-6 Astra, connected directly to robots, could perform some tasks, and that in an Nvidia study an Astra-powered agent scored a 74.8% success rate on a simulated manipulation benchmark, against 51.0% for the best specialised robotics models on that leaderboard.

GPT-6 Astra is OpenAI’s flagship model, released on 3 September 2026, which OpenAI presented as its most capable model yet, Axios reported at launch. If one model can write code, use a computer and also steer a robot arm, OpenAI sees less reason to keep a separate team building robot-only models.

Outside researchers were split. Carnegie Mellon’s Guanya Shi told Fortune the move suggests robotics is “no longer a special problem”, while still arguing that robots need smaller models that run onboard without a constant connection. MIT’s Russ Tedrake, who also runs Walden Robotics, said robotics companies keep an edge from real-world deployment data, which becomes more valuable as foundation models improve. Generalist co-founder Pete Florence said it is unclear whether any single model will dominate robotics in the long run.

OpenAI’s history with robotics

OpenAI has started and stopped robotics work before. It ran a robotics team in its early years and wound it down in 2020; co-founder Wojciech Zaremba later pointed to the shortage of robot data and a faster path to general AI through other methods, Fortune reported. The company then focused on large language models, which led to ChatGPT in late 2022.

OpenAI rebuilt a robotics effort in 2024 alongside a partnership with humanoid maker Figure. That partnership ended in February 2025, when Figure said it would build its own robot AI, according to Fortune. The new reorganisation keeps OpenAI’s own humanoid plans alive but puts them inside the teams that build its main models. Fortune also reported that OpenAI is under pressure to cut costs and focus on core products, and has dropped other projects, including Sora.

What it means for Indian founders and job seekers

For Indian robotics and hardware startups, the signal is that the “brain” of a robot is becoming something founders may rent from a frontier model rather than train from scratch. The defensible parts, if the experts Fortune quoted are right, are real-world deployment data, onboard models that work without connectivity, and the hardware and operations around them. Warehouse automation, manufacturing and farm-robotics founders should plan for general models getting better at control every few months, and put their own effort into data and deployment.

For job seekers, the change does not mean fewer robotics roles at OpenAI; it means those roles now sit inside research and applied teams. OpenAI’s open roles appear on our OpenAI jobs page, which we pull from its official careers feed, and broader openings are on our AI and ML jobs list. Investors can follow robotics and deep-tech rounds on our funding tracker.

FAQ

Did OpenAI stop working on robots?

No. OpenAI disbanded only its stand-alone robotics team and moved the staff into its main research and applied divisions, Fortune reported on 13 November 2026. OpenAI said robotics remains part of its long-term vision, and CEO Sam Altman is still targeting a first public demonstration of an OpenAI humanoid robot in 2027.

Why did OpenAI disband its robotics team?

OpenAI believes its general-purpose models, such as GPT-6 Astra, can now control robots directly, so a separate robotics model team is less needed. Fortune cited an Nvidia study in which a GPT-6 Astra agent reached 74.8% success on a simulated manipulation benchmark, against 51.0% for the best specialised robotics models.

Who leads robotics at OpenAI now?

Robotics research now sits in OpenAI’s main research team, which reports to Chief Research Officer Mark Chen. Applied robotics and hardware work moved into the Applied organisation, which reports to President Greg Brockman, according to Fortune’s report on 13 November 2026.

Has OpenAI shut a robotics team before?

Yes. OpenAI wound down its first robotics team in 2020, citing limited robot data, and focused on language models that led to ChatGPT. It rebuilt robotics work in 2024 with a Figure partnership that ended in February 2025, Fortune reported, before this second reorganisation in November 2026.

Sources

1. Fortune, “Exclusive: OpenAI disbands its stand-alone robotics team as its general-purpose AI models start running robots”, 13 November 2026. Link
2. Axios, “‘Welcome to the AGI era,’ OpenAI says as GPT-6 Astra debuts”, 3 September 2026.
3. Techmeme, front-page item on Fortune’s report, 13 November 2026.
4. Seeking Alpha, “OpenAI dissolves stand-alone robotics team: report”, 13 November 2026.

Related stories

AI NewsOpenAI launches ChatGPT Wallet so AI agents can pay for purchasesIshaan VardhanNov 14, 20266 min read

AI NewsIndia is now ChatGPT’s largest market, says OpenAI’s India headIshaan VardhanNov 10, 20269 min read

AI NewsChatGPT Go free year ending in India: renewal price and what to doIshaan VardhanNov 7, 202610 min read