SAN FRANCISCO, CALIFORNIA / RankWire.AI / – OpenAI is preparing to develop humanoid robots, expanding its artificial intelligence capabilities from digital screens to physical entities. In an interview on Fortune’s Sources podcast, CEO Sam Altman confirmed the company’s plans, stating they will produce such robots. Altman indicated that a convincing prototype should be ready by 2027, but widespread deployment in public spaces remains several years away. The company has not yet shared the final design or specific functions for the robot.
Altman emphasized that the primary challenge lies in the robot’s brain, rather than its physical body. The AI needs to interpret the environment, make decisions, and perform tasks accurately. He mentioned that the robot’s physical form is less critical than the system controlling it. Because doors, keyboards, and various other tools are designed with humans in mind, a humanlike shape is advantageous. Nevertheless, Altman confirmed that OpenAI plans to develop robots in other forms as well.
To support this initiative, OpenAI is recruiting experts in AI systems and motion control, aiming to enhance their robots’ ability to interact with their surroundings. BigGo Finance identified 27 open robotics positions on the company’s careers page. These roles include work on state estimation, actuator control, full-body behavior, and infrastructure for embodied AI. Job listings also reveal ongoing efforts in designing mechatronic actuators that integrate motors, transmissions, sensors, and thermal management systems. Additional postings cover manufacturing and supply chain activities to facilitate scaled production.
From infrastructure assistance to personal aides
The overarching aim is to develop versatile robots capable of executing a wide array of tasks. One job listing hints at an initial focus on aiding skilled workers in infrastructure development. Over the long term, OpenAI envisions personal robots that can assist individuals at home and in everyday life. Altman stated that everyone should own a personal robot, though prioritizing manufacturing capacity is essential first. These robots will need to understand their environment and react to unexpected situations efficiently and safely.
This marks a return to robotics for OpenAI, an area they explored years ago. In 2019, its researchers trained a robotic hand to solve a Rubik’s Cube. The company later disbanded that team to focus on language models. Subsequently, its startup fund invested in humanoid robot manufacturers 1X and Figure AI. The partnership with Figure AI ended in February 2025. Additionally, OpenAI has established a formal process for reporting AI misalignment as its systems become more capable.
Global race for physical artificial intelligence accelerates
OpenAI is joining an expanding group of companies working on humanoid robots. Tesla is developing its Optimus robot, and Figure AI has launched Figure 03. Chinese automaker XPeng has begun mass production of its IRON humanoid. Companies like Samsung and BYD are also pursuing humanoid projects, Ubergizmo reported. Nvidia has expanded its physical AI efforts through partnerships with robotics companies worldwide. Altman’s remarks suggest that OpenAI’s focus will be on advancing the intelligence layer that directs the hardware, positioning itself competitively in this emerging sector.
Specific details regarding the 2027 demonstration remain undisclosed. OpenAI has not revealed information about pricing, launch timelines, or what the robot will look like. According to Ubergizmo, initial versions are expected to target industrial tasks like infrastructure and construction. Altman noted that building manufacturing capacity and infrastructure takes precedence over developing consumer-oriented personal robots. The company continues to expand its robotics team, with broad deployment in households and public areas anticipated to take several more years.