In a move that screams “we’ll just do it ourselves, then,” OpenAI CEO Sam Altman has announced the company is officially back in the robotics business. The new division, aptly named OpenAI Robotics, is now on a hiring spree for top-tier engineers to build robots that can “help people in the physical world.” This announcement comes just a few months after the very public, very sudden end of its partnership with humanoid developer Figure.
Leading this ambitious reboot is Aditya Ramesh, a name that should ring a bell for anyone who’s turned text into a photorealistic image of an avocado-shaped armchair. As a primary creator of DALL-E and a leader on the Sora video-generation team, Ramesh’s appointment as VP of Robotics is a massive tell. OpenAI isn’t just bolting a chatbot onto a metal skeleton; it’s betting its entire generative AI and world simulation expertise can crack the code of embodied intelligence. Ramesh confirmed his new role, stating his goal is to “bring the intelligence of our video generation models to the physical world.”
Of course, this isn’t OpenAI’s first robot rodeo. Veterans of the industry will recall OpenAI’s original robotics team, famous for teaching a robot hand to solve a Rubik’s Cube, which was unceremoniously disbanded in 2021. At the time, the company cited a lack of high-quality training data as a major roadblock. The implication now is that with powerful world models like Sora, they can simulate reality so effectively that the data bottleneck is no longer a deal-breaker.
Why is this important?
This isn’t just a new R&D project; it’s a strategic pivot that reshapes the AI landscape. After its brief but highly publicized collaboration with Figure ended—with Figure claiming it had its own “major breakthrough” and needed to vertically integrate—OpenAI is clearly done outsourcing its physical-world ambitions. This move puts OpenAI in direct competition not only with its former partner Translation not available (en-us) but with every other company in the increasingly crowded humanoid space.
By placing a generative model guru at the helm, OpenAI is signaling a fundamental belief: that the hardest part of robotics isn’t the hardware, but the brain. Their bet is that a truly intelligent, world-aware AI can overcome mechanical limitations, a philosophy that will now be put to the ultimate test in the unforgiving realm of physical reality.