OpenAI advertises robotics roles with salaries up to $500,000

OpenAI is expanding its robotics operation, advertising 27 positions with base salaries ranging from $177,000 to $500,000 a year, according to a review by Business Insider. The figures exclude equity and apply to roles across engineering, laboratory work, data operations and legal support.

The number of robotics openings on OpenAI’s careers page has risen from 11 in May to 27, Business Insider reported. Taken together, the listings suggest the company is building capabilities beyond the software that directs robots. It is also seeking people who can design machines, assemble and test them, and produce the real-world data required to improve their performance.

A robotics operation taking shape

OpenAI is recruiting actuator design engineers, whose work involves the motors and mechanisms that move robotic joints. Other vacancies cover software, firmware and machine learning. A laboratory technician would help create and test robotic systems, while a lawyer would work specifically with the robotics group.

One position would manage the company’s data-collection facilities. That focus points to a major challenge in robotics: useful training material cannot be gathered from the internet as easily as the text and images used to develop many AI systems. Robots need examples involving physical surroundings, human movements and machine interactions. Companies therefore have to create much of that information themselves or obtain it from specialist suppliers.

The largest advertised salary belongs to a machine-learning engineer responsible for distributed data systems. The job concerns the infrastructure used to process and transfer large volumes of robotics training material among computers, with a listed base salary of as much as $500,000.

Guy Hoffman, a Cornell University professor who leads its Human-Robot Collaboration and Companionship Lab, examined the 27 listings for Business Insider. He said the recruitment drive appeared to point to a team developing robots designed specifically for OpenAI rather than merely adapting existing machines.

The advertisements describe ambitions for general-purpose robots and mention work across different physical designs. One listing outlines a future in which people could have a personal robot capable of handling whatever tasks they require. The postings do not, however, identify a final product or explain when one might be released. OpenAI declined to comment to Business Insider.

From AI models to machines in the world

The hiring push comes as investment increases in physical AI, a term used for systems that allow advanced AI models to operate in the physical environment. OpenAI’s chief executive, Sam Altman, has become more direct about the company’s plans. In an interview with investor Alex Heath on the Sources podcast earlier this month, Altman said OpenAI would make a humanoid robot while also pursuing other designs.

The robotics team is led by Aditya Ramesh, a researcher associated with DALL·E, OpenAI’s image-generation system, and Sora, its video model. Ramesh has also worked on the company’s world-simulation models, which aim to represent how the physical world behaves.

OpenAI has tried robotics before. The company ended a project in 2020 that had produced a robotic hand capable of solving a Rubik’s Cube. Its renewed effort follows Business Insider reporting earlier this year that OpenAI was training a robotic arm on household activities as part of work connected to humanoid machines.

Altman has also pointed to data centers as an early setting for robots, with household assistance as a longer-term possibility. If OpenAI develops a humanoid system, it would enter a field that already includes heavily financed efforts by Tesla, which is working on Optimus, and Figure AI. OpenAI’s affiliated OpenAI Startup Fund has invested in Figure, and OpenAI has previously collaborated with the company on robot-focused AI models.

What the listings reveal—and leave unclear

The vacancies offer only limited clues about the machines themselves. References to batteries and laser range finders led Hoffman to suggest that OpenAI may be considering a mobile robot that can operate without a fixed connection. Those details do not establish whether it would travel on wheels or legs.

Hoffman also noted that the listings seek relatively little electronics expertise. That could mean the company is not making advanced sensor hardware a central priority, instead relying on standard cameras, microphones and laser range finders. These devices can help a robot perceive its surroundings, measure distances and navigate.

For now, OpenAI’s recruitment provides a clearer picture of the infrastructure behind its robotics plans than of the robots’ eventual form. The breadth of the jobs indicates an effort spanning physical design, testing, data generation and machine-learning systems—areas the company will need to combine if it is to move from experimental robot software to machines that can work reliably in homes, workplaces or data centers.