OPENAI chief executive Sam Altman has publicly confirmed that the company is developing its own humanoid robot, a move that has long circulated in industry chatter. Speaking on the Sources podcast, Altman reiterated that OpenAI intends to pursue a robot in a human form, arguing that human environments and tools are designed for people, and therefore a humanoid shape makes practical sense for navigating real-world spaces. When pressed about whether OpenAI would build the robot in-house, he answered succinctly: “We will.”
The conversation emphasised that the real technical challenge lies in the robot’s brain rather than its appearance. Altman described the need for a capable, integrated system that can perceive its surroundings through cameras and sensors, interpret instructions, plan movements, and coordinate complex actions, such as lifting a mug with appropriate grip and force.
He contrasted the difficulty of translating software models into reliable physical action with earlier robotics work, noting that physics often governs what can be achieved in the real world. He also indicated a gradual, staged path: a demonstration could arrive in 2027, with broader real-world deployment still years away.
OpenAI’s public materials outline ongoing hardware and manufacturing considerations, including roles in motion control, state estimation, and actuator systems, suggesting an intention towards eventual mass production. Privacy and pricing details remain undisclosed, a gap that audiences and prospective users are likely to watch closely as the project progresses.