FastCompany – Tesla’s Optimus humanoid robot is still an Elon Musk pipe dream
Amid falling car sales, Musk has been hyping his humanoid robot, Optimus, as Tesla’s next big product and a potential $10 trillion business. All signs point to the contrary.
See original article by Jesus Diaz at FastCompany
Tesla is speeding toward a head-on collision with the realities of the auto industry. Amid protests over CEO Elon Musk’s involvement with President Donald Trump’s administration, not to mention Tesla’s stagnant model line and obsolete technology, the company’s revenue for the first quarter of 2025 slid 9% year over year to $19.34 billion while global deliveries shrank 13%. The second quarter was equally dire: Tesla delivered just 384,122 cars, down 13% year over year, and took in $22.5 billion in revenue, 12% lower than the same period last year. Tesla sales have now been eclipsed by China’s BYD for three consecutive quarters.
But Musk wants Tesla fanboys not to worry. Instead of focusing on the company’s declining sales figures, Musk spent most of his opening remarks on his recent earnings call talking about the enormous potential of the company’s Robotaxi service and Optimus humanoid robots to make Tesla—and its investors—very, very rich.
Those promises are still rather hollow. Tesla’s Robotaxi service, which arrived in Austin in June, is still an invite-only test limited to a small geofenced area. Competitor Waymo has more than 100 fully autonomous vehicles operating on the Uber platform in that city, covering a 90-square-mile area. Tesla covers a fraction of that in a zone shaped like a penis because Musk is a comedy genius. (“It’s going to get bigger and longer,” he told investors on his recent earnings call.) It’s also full of dangerous glitches that reveal the limitations of Musk’s self-driving solution, which currently relies heavily on tele-operators to avoid fatal accidents.
Weeks after the Tesla service launched, the city of Austin says the test remains limited to 10 to 20 Model Y cars, which look nothing like the actual Robotaxi design Musk originally presented to investors. Meanwhile, Google’s Waymo and Baidu’s Apollo Go autonomous cab services continue to expand across the planet, with thousands of cars already on the streets thanks to designs that combine AI brains with multiple optical, radar, and laser sensors.
The other promise—the really big one—is Optimus, the humanoid robot that Musk has said could
one day transform his company into a $25 trillion market-cap, world-dominating empire. In Musk’s vision, legions of 5-foot-8 electric androids running Tesla’s vehicle-grade AI will ferry battery cells, haul parts carts, and stack sheet-metal blanks across the company’s gigafactories, taking on the dull, injury-prone chores no human wants.
These bots, he has said, will be working on Tesla assembly lines by the end of 2025. Next year, they’ll roam third-party warehouses and logistics hubs across North America. Within five years, Musk promises, Tesla will be making 1 million units per year. Within a decade, there could be 100 million of them in the wild as companies worldwide deploy them for stockroom restocking, parcel delivery, elder-care assistance, and other services. Conveniently, their sales would buoy Tesla’s share price just as its core EV lineup shows its vulnerabilities.
“Optimus has the potential to be north of $10 trillion in revenue, like it’s really bananas,” Musk told investors in April. “It will be the biggest product ever,” he reiterated on the July 22 earnings call.
Except, just like with Robotaxi, it doesn’t look like that’s working out as promised. At this point, it’s extremely hard to believe that Optimus will materialize in any meaningful way within the timeframe Musk anticipates. It’s even harder to believe that the program can get off the ground quickly enough to compete with companies that appear to be further ahead in developing and deploying humanoid robots—especially in China, where competitors have the advantage of fully dominating the supply chain to manufacture bots at scale.
Market analysts predict the entire humanoid robot market could reach $13 billion to $38 billion between 2030 and 2035 (making Musk’s claims of $10 trillion in annual revenue mathematically impossible, at least for the next decade). Meanwhile, the industry is still in its Macintosh era, according to Jeff Cardenas, CEO of Apptronik, which makes the Apollo 1 humanoid robot.
“This is like the early 1980s for personal computers,” Cardenas told me recently. “We’re in the early days.” That means there’s a chance for Tesla to catch up. But it also means there’s a Mars-size gap between Musk’s $25 trillion robot revolution and the hard work of actually engineering it.
ROBOT DREAMS MEET HARD REALITY
Here is the Optimus reality today: Production reportedly froze in mid-June when the bots’ overheating joints, limp wrists, and batteries that croaked before lunchtime forced a complete procurement pause. By then, only about 1,000 bots had rolled off the Fremont, California, pilot line, costing around $60,000 apiece to make. According to videos of the bots working, they were moving at less than half the speed of the humans they are meant to replace.
Optimus’s problems have prompted a redesign to improve the robots’ performance. Tesla, according to reports, is evaluating new providers of all its key failing components, a process that will further delay an already delayed schedule.According to AInvest, two sources in the Optimus supply chain claim that “Tesla had procured enough parts to produce 1,200 Optimus units and had manufactured close to 1,000.” Though Musk had claimed the company would produce 5,000 units in 2025, according to these sources, “With the suspension of parts procurement, this target is now largely unattainable.” (The Information confirmed the report.)
On his Q2 earnings call, Musk said that Tesla is now working on an Optimus 3 prototype. “We are gonna retool a bunch of things, so it’s gonna probably be prototypes of Optimus 3 [at the] end of this year and then scale production next year,” he said.
The rosary of failures may explain why Optimus’s program chief, Milan Kovac, bolted in June. Kovac left Tesla abruptly, citing family reasons (“I want to make it clear that this is the only reason,” he insisted on social media). The project is now in the hands of Tesla’s Autopilot boss, Ashok Elluswamy.
Some old-design Optimus robots are currently in Tesla factories doing limited jobs. The company has shown the robots performing specific tasks, like folding a T-shirt and moving batteries at glacial speeds. Musk also recently reposted a video of an Optimus serving popcorn at the new Tesla Diner in Hollywood, though the clip is speeded up, giving the impression that the robot functions at a normal pace.
The Jetsons-esque diner, which houses a Supercharger station, is a smokescreen, detonated the day before the company’s Q2 results and meant to distract from yet another empty promise: Musk hasn’t yet delivered batteries that load as fast as gas pumps, so he’s now inviting drivers to fill up on pancakes and burgers while waiting for their cars to charge. The popcorn-serving Optimus, meanwhile, is a prop used by the company to distract fans and investors.
A MAGNET-SIZE PROBLEM
Optimus’s problems run deep, starting with the rare-earth magnets it needs to power its motors. Without the magnets, there’s no way to make robots that can move with the speed and precision of real humans. While Tesla hasn’t disclosed the number of magnets the Optimus requires, analyst Luke Lango says it may use up to 8 pounds of magnets. (A typical humanoid robot needs more than 40 servo motors, and each of those requires up to 3.5 ounces of magnets.)
Back in March, Musk blamed Optimus’s manufacturing problems on the lack of rare-earth magnets. Beijing had canceled exports when Musk’s former best buddy declared a trade war against China. In June, China’s Ministry of Commerce resumed issuing permits for rare-earth magnet exports, putting them on a short leash while cracking down on smuggling. But this is a problem that is not going to be solved anytime soon, if ever. As long as geopolitical tensions exist, no Western company can fully rely on Chinese magnets.
The U.S. government is trying to stop Chinese dominance by investing in MP Materials, an American company that just built a small magnet supply chain that’s still in its embryonic state. To make the millions of magnets that Musk needs for Optimus to succeed at scale, Tesla would have to put itself at the front of the line, ahead of everyone from carmakers to gadget peddlers, including Apple (which itself bought into the MP Materials pipe dream this month). Such a supply chain will be very hard to build and could disintegrate at any time.
DESIGNED TO FAIL?
Even if—and that is a Starship-size if—Tesla manages to secure a steady supply chain to scale up production to the levels that Musk needs, Optimus’s design may be doomed from the start.
Tesla is chasing the hardest form factor in robotics, humanoids, without lidar, which means its robots will encounter the same issues that plague the company’s Full Self-Driving mode and Robotaxi. While Musk claims that lidar is for losers, the exclusive use of regular cameras has proven insufficient to understand the world around cars, leading to accidents. (From 2019 to 2023, Tesla’s Autopilot was involved in 736 crashes, including 17 fatal ones. And that’s without counting the near-misses and car bumper accidents.) It will be the same with robots, which is why many of Tesla’s rivals, including startups like Unitree and AgiBot, are leaning on proven sensor stacks.
And even then, the humanoid form factor may not be right in some settings. Chris Walti, Tesla’s first Optimus lead, who left in 2022 to found the non-humanoid robotics startup Mytra, recently argued that human-shaped robots are “the wrong option for factory work” because the human form “evolved to escape wolves and bears,” rather than perform repetitive industrial tasks, making it a “suboptimal system.”
The reality today is that specialized, purpose-built automation consistently outperforms general-purpose humanoid robots in cost, reliability, and efficiency for industrial tasks. That makes Tesla’s approach both an expensive engineering exercise and one that may lack a viable business strategy. Unless Optimus mutates into something tougher, easier to make, and a lot cheaper (Unitree’s model has a $16,000 price tag), it’s destined to remain a showroom tease for a long time.
Apptronik’s Cardenas is still bullish on humanoids, which he argues can one day be deployed to perform a multitude of tasks. (Most bots today are created for single purposes.) “My view is that the key thing for robotics to scale is versatility,” he says, noting his belief that we will someday see humanoids everywhere in the world, both legged and wheeled. Investors seem to agree. Financing for the sector has been growing wildly, especially in China, where the humanoid robot industry is most advanced. There, total funding reached $4.4 billion as of July 2025, doubling the previous year’s high-water mark, according to data from ITjuzi, which tracks China’s venture-capital market.
On July 21, Chinese e-commerce giant JD.com unveiled investments in three Chinese robotics startups: Engine AI, LimX Dynamics, and Spirit AI. Meanwhile, Hangzhou-based Unitree reached unicorn status in June with a new funding round that valued it at more than $1.3 billion. Investors included Alibaba Group Holding and affiliate Ant Group; Tencent Holdings; China Mobile; carmaker Geely; and Jinqiu Capital, an investment firm founded by former ByteDance employees.
Many of these Chinese companies are already lining up customers. Hong Kong-listed UBTech secured a $12.6 million purchase from state-owned car exporter MiEE (Shanghai) Automotive Technology, by far the largest deal in the humanoid industry. Unitree and AgiBot also struck deals with telecoms operator China Mobile, worth $6.4 million and $10.9 million, respectively. Unitree founder and CEO Wang Xingxing said the company’s annual revenue had surpassed $139.4 million.
These deals are far from the trillions that Musk covets, but they are a clear signal that his company is behind. Three months ago, Musk promised an army of Optimus robots marching through Tesla factories by the end of the year. Looking at the current state of the project, his 100-million robot party is turning into the Fyre Festival of robotics.
See original article by Jesus Diaz at FastCompany
