The IEEE-RAS International Conference on Humanoid Robots closed its paper submission window on July 24, and with it, the field quietly said something about itself. For its 25th edition — December 6–9 in Santa Clara — the organizing committee picked a special theme: “Humanoids and the Future of Work.” Per TechTimes, that phrase is not decoration. It’s the research community’s own admission that the interesting question about their machines is no longer whether they function, but what happens to the people whose jobs they’ll do.
A conference that used to ask if robots could walk
The first Humanoids conference met in 2000, the same year Honda unveiled ASIMO, when getting a bipedal machine to walk across a stage without falling was a genuine research triumph. For most of the 25 years since, the field’s hardest problems were mechanical and mathematical: balance, gait, dexterous hands, whole-body control. Nobody needed a “future of work” track, because the robots had no future of work — they had a future of not tipping over.
That the 25th edition formally solicits papers on “ethical and social challenges of humanoid robot deployment” as a distinct research category is the tell. When the people building the machines add a paper track about the consequences of the machines, the consequences have arrived. You don’t convene the world’s roboticists to debate labor displacement unless the labor displacement is real enough to debate.
The machines are, in the conference’s own phrase, on the clock
The theme tracks the hardware. In the run-up to this conference, humanoid robots crossed from pilots into paid shifts. Figure AI put 40 of its Figure 03 units into BMW’s Spartanburg, South Carolina plant in late June, after an 11-month Figure 02 trial that helped build more than 30,000 BMW X3 vehicles. Boston Dynamics shipped its first commercial electric Atlas units to Hyundai and to Google DeepMind. Agility Robotics’ Digit is pulling commercial shifts at a Toyota plant in Ontario and a GXO warehouse in Georgia — and on June 24 the company announced a SPAC merger valuing it at $2.5 billion, headed for Nasdaq as the first publicly traded pure-play humanoid maker with robots already at work.
Then there’s China’s volume story, which is the one that should make the “future of work” theme feel less like a panel topic and more like a countdown. AgiBot rolled out its 15,000th robot in late June — less than three months after hitting 10,000. Unitree has built roughly 11,000 bipedal units and sells its G1 at $16,000, a price that makes “should we hire or deploy” a real budget line rather than a thought experiment. China’s industry ministry told the World AI Conference it expects to produce over 100,000 humanoids this year.
Who got put in charge of the theme
The detail worth sitting with: the conference handed its “Future of Work” programming to Special Topic Chairs Gary Bolles, who writes about the future of work at Singularity University, and Nell Watson, an AI-ethics researcher. Not two locomotion specialists. The committee staffed the flagship humanoid-robotics event with the people who think about labor and ethics, and pointed them at the questions that follow deployment: who deploys these systems, under what conditions, and with what consequences for the workers already doing those jobs.
That’s a striking self-awareness for an engineering field that has spent a quarter-century mostly celebrating capability. The roboticists in that room in December will know the deployment curve better than any policymaker — AgiBot’s 1,000-to-15,000 ramp happened in months, not years, and Goldman Sachs’ oft-cited projection has the market growing roughly 13-fold by 2035. When the people with the best view of the curve name “the future of work” as the thing worth convening around, that is not marketing. It’s a weather report from inside the storm.
The LostJobs read
Conferences pick themes to signal where a field thinks it is. For 25 years, humanoid robotics signaled “we’re solving hard engineering problems.” In 2026, it’s signaling “we need to talk about jobs.” The machines moved from the demo reel to the factory floor, and the field noticed before most of the workforce did.
If you’re in logistics or automotive assembly — the two sectors where these robots are already clocking in — the useful takeaway isn’t panic; it’s timing. The engineering case is now settled in exactly the high-repetition physical roles the robots are being deployed into, and the deployment is compounding month over month. The people best positioned to tell you how fast just scheduled a four-day conference to talk about it. When the builders start holding sessions on the social fallout of their own product, that’s the signal to plan around — not the press release that says everything’s fine.
