Robots cooking and cleaning in your house. Building cars and even bigger machines in factories. Serving you burgers at a 1950s-style diner.
That’s the futuristic vision Elon Musk believes will one day drive his electric vehicle company Tesla to a stratospheric market value of $25 trillion — more than five times higher than the world’s current most valuable company.
In a post on his social network X on Monday, Musk claimed that Tesla’s Optimus humanoid robots — currently being built under enormous secrecy — will eventually account for 80 percent of the company’s revenue, even as it suffers a worldwide sales slowdown.
That’s not the only eyebrow-raising prediction Musk has made about Optimus. He insists that Tesla will be building between 500,000 and 1 million robots by the end of 2027, and has even projected that by 2040 there will be more humanoid robots on Earth than human beings.
“My prediction is … that Optimus will be the biggest product of all time by far,” the tycoon told Tesla employees at an all-hands meeting in April. “Nothing will even be close. I think it’ll be 10 times bigger than the next biggest product ever made.”
That said, there are a few teensy tiny snags. For one thing, Tesla is reportedly struggling to fulfill even Musk’s initial promise to build 5,000 Optimus units this year.
For another, Optimus’s major public outings so far have often been under at least partial remote control by human operators, rather than fully autonomous.
All of which has left robotics experts skeptical that Musk will hit his sky-high targets — especially given his long history of failed predictions and broken promises.
“Elon is going back to his standard playbook of making outrageous claims,” Prescott Watson, a venture capitalist with RedBlue Capital who specializes in vehicles and drones, told The Independent.
“Do we really think he’s going to be selling $100 billion of humanoid robots in a few years from now? Somehow that feels even more suspicious than his announcement of the new [Tesla] Roadster in 2017.”
Gary Marcus, a cognitive scientist and AI expert based in Seattle, likewise said Musk’s promises were “not impossible, but doubtful.”
“They may actually build 5,000 test robots this year, but seeing 500,000 in 2027 would be phenomenal, and seems very unlikely,” he told The Independent.
Even some people close to the project are unsure of Musk’s predictions. Back in 2023, UCLA robotics professor Dennis Hong led the creation of a humanoid robot called ARTEMIS, whose core team are now working on Optimus at Tesla.
“Tesla is indeed preparing for mass production. The robot’s locomotion has improved dramatically, and its manipulation skills are impressive,” Hong told The Independent.
But he added: “Though I share the ultimate goal for humanoid robots with Elon, I still think Elon’s timeline is too optimistic.”
‘Optimus will plant a U.S. flag on Mars’
The stakes are high for Musk’s big robotic gamble. Increasing competition from cheaper Chinese electric vehicles, as well as the PR impact of Musk’s incendiary alliance with Donald Trump, has led Tesla’s EV sales to plummet across the world.
Apparently undaunted, Musk has been talking up a new, utopian vision of “sustainable abundance”, in which customers will enjoy their own “personal robot buddy” which helps them with household tasks like babysitting and gardening.
According to tech insiders’ news site The Information, the Optimus project is so cloak-and-dagger that its hundreds of employees are invisible on Tesla’s internal org chart, with manufacturing sites and research labs locked down to other personnel.
Some Tesla backers are all in. “A fleet of Starships is going to land on Mars, and a landing ramp is going to come out of each of them, and you’re going to see Optimus robots walk down the ramp and plant the American flag,” said investor Gavin Baker at a conference held by The Information this April.
Yet early signs have been mixed. In June, Optimus engineering chief Milan Kovac quit Tesla after nine years at the company. Some almost-finished Optimus models are reportedly sitting around with no hands as Tesla struggles to perfect their manual dexterity.
‘It’s not clear humanoid robots are worth it’
Part of the problem, experts say, is that Musk’s insistence on a humanoid design risks creating a machine that is a jack of all trades and a master of none.
“It’s not even clear that the humanoid robot race is worth running,” explains Watson, the venture capitalist. He says humanoid bots can excel in unpredictable environments, where flexibility and adaptability are more important than sheer speed or efficiency.
But in predictable and specialized environments such as factories, it’s questionable whether a general purpose humanoid design can beat specialized industrial robots.
“Humanoids will be useful some of the time, doing some tasks,” says Gur Kimchi, the former head of Amazon’s drone delivery program, who is now building a competing, more specialized robot.
“But to be good all the time on all tasks is extremely challenging, and often the humanoid morphology makes it unnecessarily complex.
“Taking a page from the animal kingdom, we have many morphologies each addressing a specific ‘ecological niche’ — in other words, optimized for a different scenario.”
Optimus’s own former team leader Chris Walti has voiced similar concerns, telling Business Insider in May that aping humanity was “not a useful form factor” for industrial work.
Even people optimistic about the future of humanoid robots have doubts about Tesla’s strategy. Andra Keay, the managing director of Silicon Valley Robotics, predicts there will be 1 million of these machines in production by 2030.
But she says Tesla is not one of the companies she’s betting on due to Musk’s insistence on trying to reinvent the technology “from scratch” in his own idiosyncratic way, adding needless complications and ignoring the lessons of past successes.
“While this approach has worked for Musk at SpaceX, it didn’t work for Tesla self-driving, which he predicted would be fully autonomous from 2014,” Keay told The Independent.
“The failures of Optimus stem from poor strategy in understanding of the complexity of robotics and the optimal requirements for ‘minimum viable humanoids’, rather than in execution.”
Moreover, the A.I. required to make a robot that can do what it’s told across a wide range of mundane tasks is highly complex, especially given that a humanoid model can be dangerous if it falls over or bumps into things. According to Gary Marcus, no company has cracked this yet.
Then there is cost. Tesla would have to beat the low-priced humanoid robots that are already available from Chinese companies, with one model selling for under $6,000 compared to Musk’s promised $20,000 to $25,000.
All of which would seem to make it a very real prospect that Optimus will be Tesla’s great debacle rather than its saving grace.
Still, perhaps Musk would merely view that as a temporary setback. “I guess when you think on an interplanetary timescale, the difference between three years and 10 years doesn’t matter much,” says Watson, tongue seemingly in cheek.
“So hey, maybe we’ll have $100 billion of Optimus robots walking around in 2028 — or 2038, who cares?”
Tesla did not immediately respond to a request for comment.
