SAN FRANCISCO, CALIFORNIA / RankWire.AI / – OpenAI is preparing to develop humanoid robots, extending its artificial intelligence capabilities from digital screens into tangible physical entities. CEO Sam Altman confirmed this initiative during Fortune’s Sources podcast, where he told host Alyson Shontell that the company intends to produce such robots. He mentioned that a convincing prototype for demonstration should be ready by 2027. Widespread deployment in public areas remains several years away. The company has not yet disclosed the final design or the specific functions these robots will serve.

Altman emphasized that the greatest challenge lies in the robot’s cognitive system rather than its physical form. The AI must interpret the environment accurately, make decisions, and execute tasks with precision. According to him, the shape of the robot is considerably less important than the system controlling it. While a humanlike form is practical because of existing infrastructure like doors and keyboards, Altman also confirmed that OpenAI plans to develop robots with different body types beyond humanoids.
To support the development of these robots, OpenAI is recruiting experts in AI systems and motion control. According to BigGo Finance, the company’s careers page currently lists 27 open positions related to robotics. These roles include work on state estimation, actuator control, full-body behavior, and machine learning infrastructure for embodied AI. Job advertisements indicate efforts to design mechatronic actuators that integrate motors, transmissions, sensors, and thermal management systems. Additional openings focus on manufacturing and supply chain processes necessary for scaled production.
From infrastructure assistance to personal companions
OpenAI’s vision involves creating general-purpose robots capable of performing a wide array of tasks. One of the positions highlights an initial focus on aiding skilled workers with infrastructure construction. Over the long term, the company envisions personal robots that can assist users at home and in everyday activities. Altman expressed the belief that everyone should have a personal robot, but emphasized that manufacturing capacity is the priority first. These robots will need to be capable of understanding their surroundings and managing unexpected situations safely and efficiently.
This renewed focus on robotics marks a return to an area that OpenAI explored several years ago. In 2019, researchers trained a robotic hand to solve a Rubik’s Cube. Later, the company disbanded that team to focus on language models. Its startup investment fund then backed humanoid robot firms 1X and Figure AI, though the partnership with Figure AI ended in February 2025. OpenAI has also established a formal process for reporting AI misalignment as its systems advance in capability.
Growing competition in physical AI development
OpenAI joins an expanding group of organizations working on humanoid robots. Tesla is developing its Optimus robot, while Figure AI has introduced Figure 03. Chinese automaker XPeng has begun producing its IRON humanoid, and companies like Samsung and BYD are also pursuing similar projects, according to Ubergizmo. Nvidia is expanding its physical AI initiatives through collaborations with global robotics partners. Altman’s remarks suggest OpenAI’s strategy will focus on enhancing the intelligence layer that directs the hardware, positioning itself to compete effectively in this space.
Specific details regarding the 2027 demonstration remain undisclosed. The company has not revealed information about pricing, launch timelines, or the robot’s appearance. Initial models are expected to be aimed at industrial infrastructure and construction, as reported by Ubergizmo. Altman stated that manufacturing and infrastructure development will precede the deployment of personal robots for consumers. The company continues hiring across its robotics division, with full integration into homes and public environments anticipated to take several more years.
