'We will have a cool demo in 2027': What Sam Altman said about OpenAI's humanoid robot and the challenges

As the artificial intelligence (AI) race heats up, with companies and governments competing for dominance, OpenAI is reportedly looking beyond AI models and consumer hardware. CEO Sam Altman has indicated that the company plans to build a humanoid robot.

According to Engadget, the OpenAI chief explained that a humanoid design makes sense because "the world is very much designed for people." Everyday objects such as doors, keyboards, and tools are built for human use, making a similar body shape a practical choice for robots navigating those environments. In his words, “matching that form factor seems good.”

When asked whether OpenAI would actually build a humanoid robot, Altman said, "We will." However, his priority is the intelligence needed to operate the robot rather than its physical design. According to Altman, questions about the humanoid robot's appearance are "less important than really figuring out ... the brain that makes the robot work." Giving a machine arms and legs will not be enough if it does not know what to do with them.

While Altman has not specified when the robot might become a commercial product, he told Fortune, "I think we'll have a cool demo in 2027 ... I don't think there will be robots walking the streets for a few years longer." The ChatGPT maker has yet to reveal what the demonstration will include, what the humanoid robot will look like, or when people outside the company might be able to access it.

The challenge of making a humanoid work: How OpenAI is tackling it

For a humanoid robot to perform everyday tasks, it must process information from cameras and sensors, understand instructions, plan its actions, and coordinate its movements. It also needs to adapt when conditions change. For instance, picking up a mug requires the robot to recognise the object, locate its handle, and apply enough pressure to lift it without losing its grip or spilling the contents.

OpenAI is hiring specialists to tackle these challenges. Its robotics roles cover motion control, estimating a robot’s position and movement, controlling actuators and coordinating whole-body behaviour. Other roles focus on building the machine-learning infrastructure needed to train AI systems for physical tasks.

The company's earlier research provides some background to this effort. Through its Dactyl project, OpenAI trained a robotic hand entirely in simulation before using it to manipulate real objects. It later used the same approach to solve a Rubik's Cube. These experiments highlighted a persistent challenge in robotics: translating skills learned in simulation into reliable actions in the physical world.

OpenAI’s hardware plans yet to take shape

The report notes that Altman’s emphasis on the robot’s intelligence does not mean OpenAI is overlooking its physical design. Its robotics job listings point to work on custom electromechanical actuators incorporating motors, transmissions, sensors, and thermal systems. Other openings focus on manufacturing and supply chains, suggesting the company is also preparing for the demands of producing hardware at scale.

For OpenAI, the challenge will be bringing intelligence and hardware together in a robot that can reliably handle everyday tasks, a goal that remains some distance from becoming a commercial reality.