‘It’s coming’: Sam Altman warns humanoid robots could soon take over human jobs

In a candid and wide-ranging conversation on Bloomberg’s The Circuit with Emily Chang, OpenAI CEO Sam Altman delivered a stark prediction: humanoid robots are coming—and they will take over jobs. While artificial intelligence is already disrupting traditional work structures, Altman suggests that the arrival of physical AI, in the form of autonomous humanoid machines, could accelerate job displacement faster than ever witnessed before.

“I don’t think the world has really had the humanoid robots moment yet, and I don’t think that’s very far away,” Altman said during the interview. “What happens when the humanoid robots get here? I mean, obviously do a lot of jobs.”

This marks one of Altman’s clearest public statements on the convergence of AI software and robotics, hinting at a seismic shift in both employment and industry.

Altman’s central argument isn’t just that AI will change work—it’s the speed with which it will do so.

“If you look at the history of the world, technological-driven job change—when one class of jobs goes away and another one pops up—that’s very consistent,” he said. “But the thing that is different this time is just the rate with which it looks like it will happen.”

Altman believes that, while AI will eliminate some roles, it will also create many new ones, particularly in engineering, science, and infrastructure. But he warned that society may struggle to keep up with the adjustment.

Stargate: The $500 Billion AI Infrastructure Push

The interview also delved into Stargate, OpenAI’s ambitious infrastructure partnership with Microsoft, Nvidia, Oracle, Arm, SoftBank, and others. First teased as an internal codename, Stargate is a $500 billion mega-project intended to build out next-generation AI infrastructure—including data centres, energy grids, and compute clusters across the United States.

“We used to think a lot about the compute we would need to train the models,” Altman noted. “What we didn’t used to think about was how much people were going to use these models.”

The first data centre under Stargate is already being built in Texas and will serve OpenAI’s upcoming models.

Altman linked the surge in AI usage—fueled by trends like AI-generated Ghibli-style images—to the rising demand for GPUs. The infrastructure being built now, he said, will help meet future spikes in inference workloads.

When pressed on the types of jobs at risk, Altman didn’t offer a definitive list, but the implication was clear: manual and repetitive roles—especially those that can be mimicked by AI agents or humanoid robots—may disappear.

He added that by 2025, agents will be doing most of the work we already know how to automate, and that 2026 could be a breakthrough year in scientific discovery led by AI.

Altman also responded to questions about DeepSeek, a rival AI firm allegedly developing more efficient LLMs.

“I think the DeepSeek team is very talented... but I don’t think they figured out something way more efficient than what we figured out,” Altman said.

OpenAI has accused DeepSeek of copying its model architecture via distillation, a controversial practice in the AI world.

Altman acknowledged the importance of better infrastructure to staying competitive. “We will have better chips. We will have better energy sources. We’ll have better algorithms,” he said. “We will optimise.”

On Microsoft and Trump’s Role

When asked why OpenAI couldn't simply scale with Microsoft’s compute alone, Altman said, “We do get a lot of great stuff from Microsoft but... this is more than any one company can deliver.”

Regarding the potential second term of US President Donald Trump, Altman was unexpectedly upbeat: “I think he [Trump] will get to make some of the most important decisions anyone in the world has gotten to make related to AI,” Altman said, praising the former president’s grasp of the industry’s stakes.

What Now?

Altman’s message was clear: the AI and robotics revolution is not speculative—it’s imminent. While he remained optimistic about long-term benefits like accelerated science and economic growth, he acknowledged the uncertainty.

“Do I think I could have sat here in 1905 and told you what we were about to discover in physics… and that 40 years later we would, like, have an atomic bomb? Definitely not.”

In that spirit, Altman urges caution, imagination, and preparedness—for a future coming faster than most expect.