Elon Musk: 100 Billion Robots in 20 Years
Tesla CEO Elon Musk, in an interview with China Central Television's finance channel, predicted the future development of humanoid robots, stating, "My prediction is that there will be at least 1 billion humanoid robots in the next ten years; in 15 years, it may reach 10 billion; and in 20 years, it could reach 100 billion."
He described scenarios where everyone could have a personal robot to care for elderly parents, look after children, serve as a one-on-one tutor for kids, and perform various tasks. In his envisioned future world, a person could simultaneously control hundreds, thousands, or even tens of thousands of physical robots and digital intelligences.
This is not the first time Musk has made such predictions. On January 22 at the World Economic Forum in Davos, he stated, "My prediction is that the number of robots will exceed the number of humans," and mentioned that humanoid robots are expected to become common household products for looking after pets, children, and elderly relatives. At the G20 Innovation Ministers' Meeting in Chapel Hill, North Carolina, on September 1, he referred to the prediction of 1 billion robots in ten years as a "conservative estimate" and expressed his willingness to bet real money on this judgment.
Musk attributes the key to achieving this goal to three aspects: AI software capabilities, onboard computing power, and the dexterity of robotic hands. He believes that software is still the "weakest link" in the current robotics industry chain.
Professor Angelo Cangelosi from the University of Manchester's Machine Learning and Robotics department publicly commented that Musk's prediction is "absolutely unrealistic," pointing out that general-purpose robots will be difficult to achieve within the next 25 years, and existing robots "can only perform tasks within a specific subset, far from being truly humanoid." He also warned that publicly demonstrated videos often obscure "the hundreds of failed attempts behind each successful action" or the involvement of human remote control.
From the perspective of industrial reality and expected gaps, the current global annual shipment of humanoid robots is only about 20,000 units (2025 data), and Bank of America predicts that by 2030, the annual shipment will only be about 1.2 million units, which presents a significant gap compared to Musk's prediction of 1 billion units in ten years. The current speed of humanoid robots in China completing tasks is about 20% of that of humans. Such long-term predictions that far exceed current industrial capabilities objectively help to continuously attract capital and attention to Tesla's Optimus and other humanoid robot projects and their upstream and downstream supply chains—benefiting related concept stocks and component manufacturers, while putting pressure on the entire industry, which needs to prove mass production feasibility in the short term but is burdened with overly high expectations. As Cangelosi pointed out, the economic logic paradox is that the stronger the robot's capabilities and the more tasks a single unit can perform, the total number of required robots will theoretically decrease, which creates inherent tension with the narrative that "the number of robots will increase exponentially." Musk himself also admitted that his predictions "are wrong half the time."
According to statistics, even if we consider the robot price Musk mentioned to be around $20,000 to $25,000, the vast majority of the global population would still find it difficult to afford the cost of personal or family purchases. The widespread application of robots in households is generally estimated to require at least another 8 years.
Source: Public Information
ABAB AI Insight
Musk's ultra-long-term radical predictions have a clear historical context—since 2015, he has reiterated every year that "fully autonomous driving will be achieved next year," which has yet to be fully realized; in 2019, he predicted that "1 million robotaxis will be on the road within a year," which also did not materialize; the mass production and delivery of the Cybertruck and Tesla Semi have been delayed by several years from the initial commitments. The humanoid robot Optimus itself follows a similar rhythm—when it first appeared at the "AI Day" in 2021, it was merely an actor in a bodysuit, and it wasn't until 2022 that a real prototype completed a walking demonstration. This round of predictions has also been continuously ramped up: at the Davos Forum in January, it was a qualitative judgment that "the number of robots will exceed the number of humans," upgraded to a quantitative "conservative estimate" of "1 billion units in ten years" at the G20 meeting on September 1, and further expanded during the interview with CCTV Finance to "1 billion, 10 billion, 100 billion" across three escalating timelines, showcasing a typical pattern of incremental ramp-up.
From a funding perspective, Tesla's valuation narrative has clearly shifted in recent years from its electric vehicle core business to "AI and robotics company," with Musk repeatedly stating that Tesla's future main value will come from Optimus rather than its automotive business; against the backdrop of fierce price competition and slowing growth in the electric vehicle sector from Chinese automakers, the continuously expanding prediction of robot ownership objectively provides narrative support for Tesla to maintain a valuation multiple far above traditional automakers, while also driving capital inflow into the upstream and downstream of the robotics industry chain—servo motors, dexterous hands, sensors, and other component manufacturers—despite the returns on these investments being based on a long-term assumption that has yet to be validated by any actual shipment data.
Similar cycles of industrial hype have occurred repeatedly in the past—visions of "flying cars" in the last century and exaggerated predictions of internet user scale during the dot-com bubble attracted massive capital inflows in their early stages but were long disconnected from actual progress. In terms of industry positioning, the current stage of humanoid robots is highly similar to the early hype cycle of autonomous driving around 2015—Tesla, Figure, Boston Dynamics under Hyundai, and several Chinese companies like Yushu Technology are all making long-term shipment volume predictions that far exceed current capacity, while the actual global shipment volume in 2025 is only about 20,000 units, creating a significant gap with the narratives of "1 billion" and "10 billion".
The structural change behind this is essentially a transfer of pricing power—in the absence of verifiable near-term delivery results, the ultra-long-term numerical predictions publicly provided by the founders themselves become a tool for capital allocation, shifting the basis for company valuation from currently verifiable revenue and shipment data to the distant market space depicted by the founders' personal narratives. The mechanism at play is that humanoid robots lack historical data for large-scale commercial use to validate, making these predictions spanning ten, fifteen, or twenty years difficult to be falsified within any election cycle or financial reporting cycle, thus serving as a durable narrative asset to support valuations long-term, unaffected by the actual shipment volume gaps in intermediate years; each time the numbers are increased by an order of magnitude—from 1 billion to 10 billion to 100 billion—it effectively renews the attention and capital patience for this narrative, which is a replicable strategy in any technology field where the actual delivery progress lags severely behind verbal timelines.
ABAB News · Cognitive Laws
- The further the prediction stretches, the harder it is to falsify.
- The bigger the story, the more sustainable the valuation.
- The more zeros in the numbers, the more rounds the story can be told.