OpenAI has officially restarted its robotics business, marking a dramatic return to physical AI after abandoning the field in 2020 to focus on large language models. On June 1st, CEO Sam Altman announced the formation of a new team called "OpenAI Robotics" and began recruiting full-stack hardware, operations, systems, and machine learning engineers to "program and build robots that are truly useful to society".

Why Is OpenAI Returning to Robotics Now?

The decision stems from rapid progress on an internal research project called "Worldsim," which evolved into the robotics initiative over the past year. The project is led by Aditya Ramesh, vice president of research at OpenAI and a core developer of DALL-E (a text-to-image model) and Sora (a video-generation model). This leadership choice signals that OpenAI views robotics as an extension of its AI foundation work rather than a separate hardware business.

Timing matters here. OpenAI is preparing for a potential initial public offering, with reports indicating the company secretly submitted a draft IPO prospectus on May 22nd and plans to go public as early as September 2026. In its latest funding round completed in March 2026, OpenAI's valuation reached $852 billion, with some institutions predicting it could exceed $1 trillion at listing. Robotics gives investors a compelling narrative: a company expanding from pure software into embodied AI, bridging the virtual and physical worlds.

What Are OpenAI's Short-Term and Long-Term Robotics Goals?

According to Altman's announcement, OpenAI's robotics strategy operates on two timelines. In the short term, the company will focus on developing robots that can assist skilled workers in building future infrastructure. Long-term, the vision is more ambitious: everyone will eventually have access to a personal robot capable of meeting various needs.

This approach differs from competitors like Figure AI, which emphasizes end-to-end hardware-software integration. OpenAI's strategy follows what industry observers call "build the brain first, then grow the body," meaning the company will develop powerful world models that understand physical laws, then transfer those capabilities to robotic systems.

How OpenAI's Robotics History Shapes Its Return

• Early Leadership (2016-2019): OpenAI developed OpenAI Gym, a reinforcement learning benchmark environment, and Roboschool, an open-source robot simulation platform. The company also created Dactyl, a dexterous robotic hand that solved a Rubik's Cube using reinforcement learning and automatic domain randomization technology.
• Strategic Pivot (2020): Facing scarce robot training data and slow iteration cycles, OpenAI disbanded its robotics team to concentrate resources on large language models. This decision proved transformative, ultimately leading to ChatGPT and the generative AI boom.
• Diversified Investments (2021-2025): Rather than abandoning robotics entirely, OpenAI invested in startups including 1X Technologies (a Norwegian humanoid robotics company), Figure AI (an American robotics star), and Physical Intelligence. The most notable partnership was with Figure AI in February 2024, when OpenAI invested in a $675 million Series B and developed a dedicated multimodal AI model for Figure's humanoid robots.

That Figure partnership, however, lasted less than a year. In February 2025, Figure AI founder Brett Adcock announced the termination of cooperation, citing divergence in technical routes. Figure believed that general large models could not meet hardware requirements and that vertical integration was necessary. This breakdown prompted OpenAI to resurrect its internal robotics team.

What Financial Pressures Drive This Move?

OpenAI faces significant financial headwinds that make the robotics pivot strategically important. The company expects to incur a loss of approximately $14 billion in 2026, with cash consumption expected to increase further. The company is not projected to achieve positive cash flow until 2030 at the earliest. Its gross profit margin sits at only about 33 percent, with high inference costs eroding profitability.

Robotics offers a potential new revenue stream and a compelling growth narrative for public markets. By positioning itself as a company that can move beyond software into hardware and physical-world applications, OpenAI can address investor concerns about the sustainability of its business model and justify its trillion-dollar valuation.

The robotics restart also reflects OpenAI's confidence in its AI capabilities. If the company's world models can successfully transfer to physical robots, it could reshape how the robotics industry approaches research and development, moving away from hardware-first approaches toward AI-first design.