SAN FRANCISCO, CALIFORNIA / RankWire.AI / – OpenAI is venturing into the realm of humanoid robotics as part of its effort to embed artificial intelligence into physical devices. In an interview on Fortune’s Sources podcast, Chief Executive Sam Altman confirmed the initiative, stating that the company aims to showcase a compelling prototype by 2027. According to Altman, robots navigating public streets are still several years away. The company has yet to reveal the final design or specific functions of the robot.

Altman emphasized that the primary hurdle lies in creating the robot’s core intelligence. This system must be capable of understanding the environment, making decisions, and executing tasks with precision. The physical appearance, such as form factor, is less critical than the machine’s underlying intelligence, Altman explained. He noted that a humanlike design is ideal for environments tailored to humans, including doors, keyboards, and equipment. Additionally, OpenAI plans to develop robots in various other configurations, according to Altman.
The firm is actively recruiting experts in AI systems and motion control to develop robots capable of interacting with the physical world. According to BigGo Finance, its careers page features 27 open robotics positions. These roles encompass motion control, state estimation, actuator control, full-body behavior, and the infrastructure needed for training embodied AI. Job listings reveal that the team is creating mechatronic actuators internally, while other openings focus on manufacturing processes and supply chain logistics for large-scale hardware production.
Building versatile robots for everyday use
OpenAI envisions creating general-purpose robots capable of a broad spectrum of functions. A recent job posting highlights an immediate goal of developing robots that can assist skilled workers in infrastructure projects. The longer-term aspiration involves personal robots designed to support individuals in their daily routines. Altman expressed the belief that everyone should have a personal robot, but achieving this goal requires systems that can comprehend their environment and adapt to unforeseen circumstances effectively and securely.
The company’s robotics development extends OpenAI’s efforts beyond software and AI infrastructure. The interest in humanoid robots is not a new direction for OpenAI. Its researchers previously trained a robotic hand to solve a Rubik’s Cube in 2019, before disbanding that robotics team. Since then, the startup fund has invested in humanoid firms such as 1X and Figure AI. Notably, Figure AI ended its collaboration with OpenAI in February 2025 to concentrate on developing its own AI models.
Growing competition in the physical AI sector
OpenAI is entering a highly competitive landscape of companies working on humanoid robots. Tesla is working on its Optimus robot, while Figure AI has introduced its Figure 03 model. Chinese automaker XPeng has started producing its IRON humanoid, and Nvidia has expanded its physical AI work through collaborations with robotics partners worldwide. Altman’s remarks suggest that OpenAI intends to distinguish itself by focusing on the intelligence that directs the hardware, rather than solely on the physical design.
The company has not disclosed details about what its 2027 demonstration will entail or the appearance of the robot. Information regarding pricing or the timeline for product launch remains unavailable. Altman stated that manufacturing capacity and infrastructure are the top priorities at this stage, with personal robots not being the immediate focus. Meanwhile, the company continues to hire engineers for its robotics division. Widespread use of robots in homes and public spaces is expected to take several more years to realize.
