OpenAI's Humanoid Robot: Home Use Isn't the Goal—Yet
OpenAI's CEO Sam Altman has finally put the rumors to rest: the company is indeed building humanoid robots. But before you start dreaming of a robot butler, he's quick to clarify that home use isn't the top priority right now.
In a recent podcast called "Sources," Altman confirmed that OpenAI is exploring two distinct paths: humanoid robots and specialized machines designed for data centers. "We definitely will develop humanoid robots," he said, but he also stressed that not every robot needs to look like us. When a human form doesn't add practical value, OpenAI will craft non-anthropomorphic robots tailored to specific tasks—like those in data centers.
So why bother with humanoids at all? Altman's reasoning is simple: our world is built for humans. Doors, computers, vehicles, even household tools—all are designed around our bodies. A robot that can navigate that environment seamlessly needs to share our shape. That's where the humanoid form becomes not just a novelty, but a necessity.
Interestingly, Altman insists that the form isn't what matters most. "The real magic is in the robot brain," he said. And OpenAI has just unveiled what they call their most intelligent and aligned model yet: GPT-6 Astra. This "super brain" could be the key to making robots truly useful.
Building the Brains and the Brawn
While Altman talks a big game about software, OpenAI isn't neglecting the hardware side. They've been quietly building a robotics team since 2024, when they first integrated their models with Figure AI's Figure 01 robot. In May of this year, they officially launched OpenAI Robotics, led by Aditya Ramesh—the mastermind behind Sora.
The team is growing fast. Job postings in the San Francisco area have jumped from 11 to 19 positions, with a heavy focus on hardware actuators. They're looking for experts in actuator design, electromagnetic systems, gear mechanics, and testing. One core role involves taking a custom electromechanical actuator from concept to mass production—balancing torque density with real-world manufacturability.
What's Next?
Altman is candid about the timeline: don't expect a robot in your living room anytime soon. The immediate focus is on infrastructure and manufacturing. But he's optimistic about the long run. "One day, I think everyone should have a personal robot," he mused.
For now, OpenAI hasn't revealed shipping dates or production targets. But with the brain and the brawn both in development, the future of robotics is looking more tangible than ever.
Key Points:
- OpenAI is developing both humanoid and specialized robots, with data centers as a near-term focus.
- Altman emphasizes that the "robot brain" (AI) is more important than the physical form.
- GPT-6 Astra, OpenAI's latest model, is seen as a foundational "super brain" for robots.
- OpenAI Robotics team is expanding, with a hiring push in hardware actuator design.
- Home robots are a long-term goal, not an immediate priority.