OpenAI officially establishes a robotics team, entering the physical world of AI.

OpenAI no longer just wants AI to exist on screens—it wants robots to enter the real world. On June 1, OpenAI co-founder Greg Brockman announced on X that the “OpenAI Robotics” team is advancing rapidly, aiming to “build AI that can help humans in the physical world” and is publicly recruiting full-stack hardware, operations, systems, and machine learning engineers. This is OpenAI’s most explicit public statement about its robotics business so far. Earlier on the same day, Sam Altman posted that OpenAI wants to “program and manufacture robots that are useful for society,” positioning robotics as a recent strategic priority. Recent goal: Assist workers, not replace them Altman clarified the direction in his post: the near-term focus is on skilled worker assistance in construction and physical infrastructure fields, rather than household or consumer-grade robots. This positioning is notable. Skilled jobs such as construction workers, plumbers, electricians have previously been considered less impacted by AI—because these jobs highly depend on manual coordination and on-site judgment. By making this sector a primary breakthrough area, OpenAI signals that automation pressure is extending to broader labor markets. Altman’s long-term vision is “everyone owning a personal robot,” but at present, no specific product, partners, or timeline have been announced. Current layout: Mainly investment, zero self-research Currently, OpenAI’s actual assets in the robotics field are quite limited. OpenAI holds minority stakes in humanoid robotics companies Figure and 1X, but has no in-house hardware. This announcement to form a team and recruit publicly is OpenAI’s first move to upgrade robotics from “investment target” to “internal strategic business.” Altman previously said on the Core Memory podcast: “We are working hard to figure out how to really succeed in robotics.” This frank statement shows OpenAI is still exploring this path. In 2024, OpenAI also established formal research collaboration with Coco Robotics, a robotics delivery startup, which is one of its few practical moves in robotics so far. Once considered splitting off, ultimately shelved This high-profile robotics strategy announcement comes with a little-known background. According to The Wall Street Journal, Altman discussed splitting off the robotics and consumer hardware departments at the end of last year, letting them raise funds and operate independently. But the plan ultimately wasn’t pursued—mainly because even if split off, these departments might still need to be incorporated into the parent company’s balance sheet, reducing the significance of splitting. Insiders at OpenAI liken the two departments to “independent startups attached under OpenAI’s roof”—they report directly to Altman rather than being included in the company’s overall management structure. OpenAI may reconsider splitting plans in the future, with one reference model being Google’s 2015 restructuring into Alphabet—placing core search alongside experimental businesses like self-driving and life sciences under the same parent company, allowing investors to assess them separately. Strategic pressure ahead of IPO The discussion of splitting is driven by bigger pressures facing OpenAI. According to The Wall Street Journal, OpenAI recently completed a $122 billion fundraising round, bringing its valuation to $852 billion. The company described it as “the largest financing in Silicon Valley history,” and expects to complete its IPO within the year. Meanwhile, OpenAI’s internal situation is not easy. The company has not met internal targets for revenue and user growth, and CFO Sarah Friar warned colleagues that slow revenue growth may affect its ability to fulfill data center commitments. OpenAI has closed its standalone Sora video app and reduced its product focus to programming tools and enterprise customers. At this point, shifting the robotics business from “considered split-off” to “core strategic priority” reflects OpenAI’s reshaping of its own narrative ahead of its IPO. Competition landscape: Not early to the game Entering robotics, OpenAI faces a batch of already established competitors. Google DeepMind, Tesla, and Boston Dynamics have all deployed scaled systems in physical AI. OpenAI’s advantage is currently in large models, but whether it can effectively convert this into robotic products remains to be seen. What OpenAI’s move signals to the industry is that foundational model capabilities may in the future be directly embedded into robotics platforms, rather than sold externally as independent APIs. For Figure, 1X, and other companies with investment ties to OpenAI, this means cooperation and competition will deepen simultaneously. Risk Disclosure and Disclaimer The market has risks and investments require caution. This article does not constitute personal investment advice and does not take into account the specific investment targets, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article are suitable for their situation. Investing accordingly is at your own risk.