Humanoid Robots Could Outnumber Humans by 2083 as Long as They Start Building Each Other

Robot-assisted manufacturing could trigger a population boom putting humanoid numbers on par with humanity.

The next population boom may have nothing to do with babies. It will likely happen on factory floors, as soon as humanoid robots begin helping to manufacture more robots in their own image.

For every million people walking the Earth today, there is roughly one human-like robot. By the end of this century, there could be many more of them than us.

A new mathematical study suggests that human-shaped robots or androids, if you will, could become as numerous as people by 2083, when the global human population is expected to approach 10.3 billion.

Using equations originally designed to describe how glass ages and how human populations surged over 12,000 years, the researchers found that conventional, human-run factories alone will likely never produce enough machines to match the human headcount. Instead, numerical parity hinges on achieving an industrial tipping point: when deployed androids begin manufacturing their own successors.

“For all of human history, we have counted people. We may soon need to count another population sharing the planet with us,” the researchers write in a statement accompanying the study.

Robots at Work

Today, the global fleet of humanoid robots is still small — numbered in the tens of thousands — but major automakers, logistics giants, and tech firms are already putting these still early units to work. In warehouse distribution centers, Agility Robotics has logged tens of thousands of commercial operating hours with its bipedal Digit robots hauling totes for Amazon, while Apptronik is piloting its Apollo humanoids with Mercedes-Benz and logistics giant GXO. On automotive assembly lines, Figure AI has deployed its robots at BMW’s Spartanburg plant, Boston Dynamics is rolling out its all-electric Atlas to Hyundai facilities, and Tesla is scaling internal factory trials for its Optimus platform.

Meanwhile, Chinese manufacturers like Unitree Robotics and AgiBot are shipping thousands of lower-cost platforms for research and industrial tasks, and companies such as 1X Technologies have opened consumer preorders for home-assistant androids like NEO designed to handle basic household chores.

The growth so far is encouraging. But even a billion robots — a huge milestone — would be vastly outnumbered by the approximately 9.7 billion people expected to inhabit Earth by the turn of this century.

×

Get smarter every day...

Stay ahead with ZME Science and subscribe.

Daily Newsletter
The science you need to know, every weekday.

Weekly Newsletter
A week in science, all in one place. Sends every Sunday.
No spam, ever. Unsubscribe anytime. Review our Privacy Policy.

Thank you! One more thing...

Please check your inbox and confirm your subscription.

To explore what might happen next, physicist Alessio Zaccone of the University of Milan, mathematical biologist Thomas Tunstall of the University of Exeter, and environmental scientist Jesse Ausubel of Rockefeller University developed a model of how human and robot populations might evolve.

Their 56-page study, published as a preprint on arXiv, adapts a mathematical framework that allows growth rates to change as populations expand.

When Robots Start Building Robots

The researchers basically looked at two scenarios. One is direct production where humans design, fund, and assemble robots in conventional factories. The other is replicative growth where installed androids actively join the industrial workforce, assembling components, managing logistics, and building new androids.

As more robots enter service, they could increase the factory’s productive capacity, potentially allowing even more robots to be manufactured. The mathematical tipping point arrives when the production system can replace robots faster than they wear out or retire. If growing robot numbers also increase productivity, the population could begin accelerating.

If this autonomous manufacturing channel switches on in 2050 with modest positive feedback, the math shifts dramatically. Android numbers surge past the 10 billion mark around 2083 — coinciding almost exactly with “Peak Human,” when the UN projects humanity will top out. From that moment on, crossover approaches from both sides: rising robot numbers meeting a shrinking human baseline.

From Thousands of Robots to Billions

How realistic is such explosive growth?

For now, the humanoid robot industry remains tiny compared with conventional industrial automation.

According to the International Federation of Robotics, approximately 7,000 full-size humanoid robots were sold worldwide in 2025 for professional and commercial applications.

But Ausubel argues that the projected scale itself is not unimaginable.

“The planet now hosts about 1.7 billion motor vehicles, and we manufacture about 100 million each year, so it is clearly feasible to populate Earth with billions of androids during this century,” he said.

Still, cars and humanoids differ substantially in their manufacturing requirements and operating demands.

The researchers tested 4,096 combinations of assumptions to see which uncertainties affected android population projections the most. Three factors repeatedly emerged: how long robots remain operational, when they start helping manufacture additional robots, and how strongly that contribution increases production.

They also examined an extreme scenario in which robot-assisted production begins in 2035 and allows the humanoid population to match humanity by 2050. Reaching that population from just 10,000 machines in 2025 through steady exponential growth would require roughly 73.5% annual growth for 25 consecutive years.

Beyond the gigantic and sustained growth required to reach parity, there are also substantial physical obstacles. Billions of humanoids would require enormous quantities of materials, semiconductors, batteries, industrial infrastructure, and electricity. The computing systems supporting them would add further demand, on top of the exponential demand currently pushed by digital AI.

The International Energy Agency already expects global data-center electricity consumption to nearly double from approximately 485 terawatt-hours in 2025 to 950 terawatt-hours in 2030. That doesn’t include the full additional energy burden of a hypothetical parallel civilization containing billions of active humanoids.

RelatedPosts

China’s New $173K ‘Biometric’ Robot Aims to be a Warm-Blooded Companion For Humans
Empowering women around the world could help usher in a sustainable society, paper argues
The Earth’s population in 2100 could be much lower than we thought
Chinese Robot Startup Releases Half-ton Mech Suit, Politely Asks People Not to Weaponize It

A New Population Question on an Old Doomsday Anniversary

The research arrives on an unusually fitting anniversary.

In 1960, physicist Heinz von Foerster and his colleagues published a paper in Science with the startling title Doomsday: Friday, 13 November, A.D. 2026.

The researchers had noticed something remarkable about human population growth over the previous 2,000 years. Not only was the population increasing, but its growth was accelerating. More people meant more potential parents, which meant even faster population growth.

They fitted a mathematical curve to that historical trend and extended it into the future. The result was bizarre, as the curve grew steeper and steeper until it shot toward infinity on November 13, 2026, which happens to be today.

Of course, the reason why you can’t model human population growth to infinity is that there are real biological limits, preventing von Foerster’s purely mathematical “Doomsday” singularity. Machines, however, play by different rules.

“Foerster’s Doomsday prediction was always bound to be thwarted by biological limits on human reproduction. Technological populations, however, will operate under vastly different constraints,” Tunstall said.

“As androids become more populous and begin contributing to the production of their successors, we may need to rethink our approach to studying them: moving beyond a perspective of android engineering towards one of android ecology.”

As Zaccone noted in the press release, the ripple effects could eventually reshape human demography itself: “A central question, once largely confined to science fiction, may eventually become a practical one for demographers, engineers, economists, and governments: not only how many robots exist, but how their presence affects human fertility, family size, and structure. I believe that studying mathematical models in this way could prepare us for a wide variety of situations, including the scenarios discussed in this paper.”