Robotics experts have criticised Elon Musk’s prediction that 10 billion humanoid robots will be in circulation by 2040, calling it “unrealistic”.
Marco Chacin, Head of Actuators and End Effectors and Robotics & R&T Roadmap Owner at Airbus Commercial Aircraft, said that even with advances in artificial intelligence and motion control, manufacturing capacity, safety regulation, and materials supply make such a scenario implausible.
At the World Economic Forum in January, Musk said he expects Tesla to begin selling its Optimus humanoid robots by the end of 2027. He billed the machines as globally transformational technology capable of performing “innumerable human tasks,” with the ambition that “you can basically ask it to do anything you’d like.” Tesla has reportedly already deployed Optimus units to perform “simple tasks” in its factories.
Musk said robots would be offered for sale only once the company is confident of “very high reliability, very high safety and a high range of functionality,” predicting that there will eventually be “more robots than people” and that “everyone on earth is going to have one and want one.”
Chacin disputed both the scale and timeline, highlighting the immense logistical, regulatory, and economic challenges of producing billions of complex machines capable of operating safely alongside humans. “We have never produced anything at that scale,” he told delegates at the Southern Manufacturing in Farnborough, Hampshire last week. “If he had said millions, maybe. Billions? Who is going to make them? Where are the materials coming from?”
The scepticism is shared by other leading figures in robotics. These include Rodney Brooks, co‑founder of iRobot and former director of MIT’s Computer Science & Artificial Intelligence Laboratory, has argued that “today’s humanoid robots will not learn how to be dexterous despite the hundreds of millions, or perhaps many billions of dollars, being donated by VCs and major tech companies to pay for their training.” Brooks also highlights that full‑size bipedal robots are not yet safe to walk among humans and predicts that while humanoids will exist in 15 years, they will look very different from both humans and today’s prototypes.
Chacin, a 26-year veteran of automation, argued that the real bottleneck for robotics is not technological progress but deployment — whether machines can operate safely and continuously in complex environments. “Don’t be fooled by the demo,” he said. “A vertical slice that works for a few minutes is not the same as an industrial system.”
He cited examples from autonomous vehicles to illustrate the gap between promise and reality. “Stanley,” the autonomous vehicle that won the DARPA Grand Challenge, navigated hundreds of miles of desert using pioneering sensors and computing, forming the foundation for Google’s early self-driving projects. Yet widespread deployment of autonomous cars stalled for years, not because of the technology, Chacin said, but because of missing infrastructure — legal frameworks, insurance models, communications networks, and regulatory approval.
Humanoid robotics has followed a similar trajectory. From early systems such as Honda’s Asimo to machines developed for the DARPA Robotics Challenge, progress has been steady but often accompanied by high-profile failures. More recently, Atlas from Boston Dynamics has demonstrated fluid, human-like motion and even the ability to swap its own battery. Such demonstrations, Chacin warned, can create unrealistic expectations about industrial readiness.
In manufacturing, the economics of robotics are shifting. Hardware is increasingly commoditised, with value migrating to software, integration, consulting, and training — a pattern Chacin compared to the iPhone ecosystem, where services and applications generate more revenue than the devices themselves. Annual global installations now approach half a million industrial robots, a figure that has accelerated markedly over the past decade.
Yet aerospace remains a marginal customer compared with automotive, electronics, and heavy machinery. The reason is structural. Even in peak years, Airbus and Boeing together have produced fewer than 2,000 aircraft annually, compared with roughly 100 million cars worldwide. Aerospace cannot justify ordering robots in the thousands, nor does it operate in tightly standardised production flows.
“In automotive, the robot is static and the chassis comes to it,” Chacin said. “In aerospace, the robot must go to the work.” Aircraft structures — 17-metre wings, 30-metre fuselages — introduce mobility, tooling, and accuracy challenges far beyond those of car plants. Safety adds another layer of complexity. Unlike automotive plants, where robots and humans can be segregated, aircraft assembly often requires close-proximity interaction.
Chacin was particularly candid about regulatory interpretation in the UK. While international ISO standards provide a baseline, he argued that inconsistent interpretation by safety officers and consultants can lead to repeated risk assessments and redesigns, slowing deployment and inflating costs. “If I deploy a system in one country, I cannot simply move it to another and expect it to pass,” he said. “Every new interpretation adds time and cost.”
Airbus has sought to shape solutions. Two years ago, the company announced a partnership with FANUC to develop a robot tailored to aerospace requirements, with greater stiffness and accuracy. The resulting platform is now being sold into other sectors.
Looking further ahead, Chacin framed robotics as a demographic necessity rather than a technological indulgence. Roughly half of global GDP is tied to human labour, yet developed economies face ageing populations and workforce shortages. Japan, where around 40% of the population is over 60, offers a preview of the pressures ahead. Robots, he argued, will enhance rather than replace human workers, freeing time for higher-value and creative activities.
“The question is not whether robotics is advancing,” Chacin said. “It is whether you can actually deploy it.”
