OpenAI has disbanded its stand-alone robotics team and moved its researchers onto other teams, Fortune reported on 13 November in an exclusive. The company confirmed the reorganization when Fortune asked about it, and a spokesperson portrayed it as a sign of OpenAI’s commitment to robotics, “rather than a retreat from it”.
The stated reason is that general-purpose models now do robot work. In an Nvidia study cited by Fortune, an agent running on GPT-6 Astra reached a 74.8% success rate on a simulated manipulation benchmark, against 51.0% for the best specialized robotic models on that leaderboard. That is a gap of 23.8 percentage points, by our arithmetic, in simulation rather than on real robots.
Business Pill · The Model and the Harness
An evergreen Business Engineer explainer on the difference between a model and the harness around it, the general idea behind giving a general model a control interface. It teaches the concept only and does not discuss OpenAI or robotics.
The key insight: As we read the Fortune report, OpenAI is treating robotics as an application of its general models rather than a separate research bet: the team is split between the main research group and the Applied organization. The evidence cited is one simulated benchmark, where an Astra agent scored 74.8% against 51.0% for the best specialized models.
OpenAI said the robotics research team now sits in its main research team, reporting to Chief Research Officer Mark Chen. The applied robotics and hardware teams moved to the larger Applied organization, which reports to President Greg Brockman. Fortune did not give a headcount.
The change was recent enough that OpenAI had already registered a booth at CoRL, a robotics conference in Austin this week. When Fortune visited booth #208, it “sat empty with OpenAI’s logo nowhere to be seen”.
Success rate on a simulated manipulation benchmark in an Nvidia study, as reported by Fortune on 13 November 2026: 74.8% for an agent running on GPT-6 Astra, 51.0% for the best specialized robotic models on that leaderboard. Simulation, not real robots; the 23.8-point gap is our arithmetic.
Why OpenAI Says It Moved
OpenAI told Fortune that at the start of the year its researchers assumed the best route to robotics was to build on models such as Sora. Since then, the spokesperson said, general-purpose frontier models including GPT-6 Astra have shown they can be applied directly to robotics. OpenAI said moving the researchers is now the best way to accelerate progress towards general-purpose robots.
The cost backdrop. Fortune also tied the move to cost pressure. It wrote that Altman told the company last year to abandon “side quests”, that OpenAI has dropped efforts including Sora, and that its operating losses reportedly totaled $10.9 billion in the most recent quarter as it seeks a new funding round and prepares for a possible IPO in early 2027.
OpenAI’s robotics team milestones as reported by Fortune on 13 November 2026. The 2027 humanoid demonstration is Altman’s target, not a scheduled event.
Fortune’s Robotics Timeline, 2020 to 2027
OpenAI wound down an earlier robotics team in 2020, Fortune said. It rebuilt a stand-alone team in 2024 alongside a partnership with Figure to develop AI models for humanoid robots. That partnership ended in February 2025, 21 months before this reorganization, by our count, when Figure said it would build its own robot AI. Fortune said Altman is targeting a first public demonstration of OpenAI’s own humanoid robot in 2027.
Guanya Shi of Carnegie Mellon told Fortune he would read the move as a signal that OpenAI sees robotics as “no longer a special problem”. He still sees a need for smaller models that run on a robot’s onboard computer without a constant connection. MIT’s Russ Tedrake, CEO of Walden Robotics, said robotics companies keep an edge in real-world deployment data. Pete Florence of Generalist said it was unclear that any existing model would dominate robotics in the long run.
The Structural Read
As we read it, the org chart now matches the claim: robotics research reports to the Chief Research Officer, and robot hardware sits with the product-facing Applied organization under Greg Brockman.
As we read Fortune’s timeline, OpenAI has now opened and closed a stand-alone robotics team twice, in 2020 and in 2026, while keeping a 2027 humanoid demonstration target. Fortune says the move also comes as OpenAI is under pressure to rein in its costs.
Guanya Shi, Carnegie Mellon University, to Fortune
“no longer a special problem”
Three Implications
FOR ROBOTICS STARTUPS Fortune says the idea that general models can run robots would seem to challenge startups that raised billions on robots needing brains of their own. Russ Tedrake told Fortune deployment data remains an advantage.
FOR OPENAI WATCHERS Fortune reports operating losses of $10.9 billion in the most recent quarter and a possible IPO in early 2027; the reorganization follows Altman’s call to drop side quests, as Fortune reported.
FOR READERS OF THE COVERAGE Fortune gives no headcount, and the 74.8% result is from simulation. Our piece on Nvidia’s GR00T 2 covers the specialized-model route.
The analysis says: “OpenAI is betting on one harness that absorbs everything.”
As we read it, folding robotics into the general research team fits that description: one model family asked to run robots too. The analysis does not discuss robotics or this reorganization.
What Is Not Established
Fortune did not report how many people were on the team or whether any left OpenAI. The 74.8% figure comes from a simulated benchmark in an Nvidia study as described by Fortune; we did not read the study itself, and simulation results may not carry over to physical robots.
The year 2024 for the rebuild is Fortune’s; it gave no month, so we do not count months to it. The $10.9 billion loss is a reported figure, not an OpenAI disclosure. We did not contact OpenAI for this piece.
This Week In The AI Supercycle
The Great Convergence of Capital, Power, and Geopolitics
This week’s issue: for two years the AI build-out was priced as a supply story, and last week as a credit story. This week it was priced as a permission story, and the answer arrived as a cost, not a ban.
OpenAI has folded its stand-alone robotics team into its research and Applied groups, Fortune reported, and the company says general models such as GPT-6 Astra now apply directly to robots. The evidence cited is a 74.8% versus 51.0% result in simulation, a 23.8-point gap. Altman’s 2027 humanoid demonstration target is the next date to watch.
A note on sourcing. This piece rests on Fortune’s exclusive of 13 November 2026 by Wen Shao, read in full, including OpenAI’s confirmation and spokesperson statements as Fortune reported them. We did not read the Nvidia study Fortune cites. The 23.8-point gap and the 21-month count are our arithmetic. We did not contact OpenAI for this piece.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
