BofA Signals Humanoids Will Outnumber Cars by 2060

Bank of America just put a number on the robot future that should make every investor and worker sit up: by 2060, the bank projects that more people will own a humanoid robot than own a car. That is not a fringe futurist talking. It is one of the largest financial institutions on Earth modeling humanoids as a mainstream asset class, and the near-term figures underneath that headline are even more revealing than the sci-fi 2060 milestone.

What Actually Happened

In a research forecast, Bank of America's analysts projected that humanoid robots will reach roughly 3 billion units in operation by 2060, a scale at which household robot ownership would exceed car ownership. The bank frames the technology as moving from laboratory novelty to a durable growth market spanning factories, logistics, services, and eventually homes, comparing the trajectory to the early years of the automobile and the personal computer.

The near-term numbers are the part that matters for anyone making decisions this decade. Bank of America estimates about 90,000 humanoid robot shipments in 2026, rising to 1.2 million by 2030. That is a roughly thirteen-fold increase in four years, the kind of compounding curve that turns a niche into an industry. The 2060 figure grabs headlines, but the 2026-to-2030 ramp is the window investors and operators can actually act on.

The forecast lands on top of a funding surge that gives it credibility. Investment in humanoid robotics climbed from $0.7 billion in 2018 to $4.3 billion in 2025, and as of early 2026 more than 50 companies are actively building humanoids, with over 150 commercial product launches already on record. The capital, the company count, and the product pipeline are no longer hypothetical, which is why a mainstream bank is willing to model the category at all.

This is also a notable shift in who is doing the forecasting. When a sell-side bank publishes adoption curves and total-addressable-market math, it is signaling to institutional clients that the sector is investable. That signal pulls a different and far larger pool of capital than the venture money that has funded humanoids so far, and it changes the conversation from whether humanoids are real to how to allocate around them.

The bank's framing also stretches the market well beyond factories. Its long-run case assumes humanoids eventually move into elder care, hospitality, retail, security, and domestic help, the high-labor service sectors where worker shortages are already chronic across developed economies. That breadth is what gets the model to billions of units; a robot confined to warehouses could never outnumber cars, but a robot that becomes a general household and service appliance plausibly could. Whether that leap from industrial tool to home appliance actually happens is the load-bearing assumption under the entire forecast.

Why This Matters More Than People Think

A forecast from Bank of America is not a neutral prediction; it is an act that helps create the outcome it describes. When a major bank models humanoids as a multi-decade growth market, pension funds, sovereign wealth funds, and asset managers start to pay attention, and capital follows attention. The presence of Brookfield, Macquarie, Intel Capital, Nvidia, and Salesforce in Figure's billion-dollar round is exactly the institutional money that bank research is built to mobilize.

The thirteen-fold shipment jump from 2026 to 2030 is where the real economic story lives. At 90,000 units a year, humanoids are a curiosity; at 1.2 million a year, they become a supply-chain category that reshapes demand for actuators, sensors, batteries, and AI compute, and a labor force large enough to move productivity statistics in specific sectors. Industries are transformed not at the visionary endpoint but during the steep middle of the adoption curve, and that middle is now forecast to begin.

For workers, the forecast reframes a question many treat as distant. If a major bank expects more than a million humanoids shipping annually within four years, the labor impact is not a 2060 problem; it is a late-2020s one, concentrated first in warehousing, logistics, and manufacturing where the early deployments already cluster. The policy debate around automation, retraining, and wages has a clock on it now, and the clock is faster than most public conversation assumes.

Capital markets also tend to front-run these curves. Public investors have already bid up anything with a credible humanoid or physical-AI story, from chipmakers supplying robot brains to industrial suppliers of actuators and reducers. A bank forecast like this one gives that trade an analytical backbone, which can pull years of expected growth into present-day valuations. The upside is faster funding for real builders; the downside is a classic setup for a hype cycle, where prices race ahead of shipments and a single disappointing quarter resets sentiment hard.

The forecast also implicitly defines the battleground. Owning more humanoids than cars by 2060 implies the home, not just the factory, becomes a market, which means consumer distribution, price, and trust matter as much as industrial capability. That favors whichever companies can drive cost down and volume up, and it explains why a $17,990 humanoid on Amazon is strategically relevant to a forecast about the year 2060.

The car comparison itself deserves scrutiny, because it is doing rhetorical work. Cars took roughly seven decades to reach near-universal ownership in rich countries, propelled by mass production, consumer credit, road networks, and a clear, singular use case: getting from one place to another. Humanoids have no equivalent killer application yet, and their value is fragmented across dozens of tasks rather than concentrated in one. Reaching car-level ownership therefore requires either a breakout consumer use case that has not yet appeared, or a collapse in price so steep that owning one becomes an impulse purchase. Bank of America is effectively betting on both, and the gap between that bet and today's reality is exactly the uncertainty investors are being asked to price.

The Competitive Landscape

The forecast maps onto a field that is already sorting into camps. On the revenue front, Agility Robotics rents its Digit to GXO under a Robots-as-a-Service contract, and Figure bills BMW around $25 per robot-hour with a factory producing one robot every 90 minutes. On the volume front, Unitree shipped more than 5,500 units in 2025 and targets up to 20,000 in 2026. On the capital front, Figure raised over $1 billion, Apptronik $520 million, and NEURA roughly $1.2 billion. The forecast's 90,000 units for 2026 will be supplied largely by Chinese volume players today.

That hands China an uncomfortable early lead in the very curve the bank is charting. Chinese firms control roughly 90 percent of current humanoid shipments, the same dominance they built in drones and electric vehicles by winning on manufacturing scale and component supply chains. If the 1.2-million-unit 2030 market is supplied mostly from Chinese factories, the financial upside that Bank of America is pitching to Western clients could accrue disproportionately to Chinese manufacturers, a tension the forecast's optimism tends to gloss over.

History offers a cautionary frame for how these curves resolve. The personal computer, the smartphone, and the electric vehicle all followed the trajectory the bank describes, but in each case the winners were decided by who controlled distribution and cost, not who demonstrated the most advanced prototype first. Japan Airlines deploying Unitree humanoids at Haneda Airport and Amazon trialing Agility's Digit are the kinds of early, unglamorous footholds that, in past cycles, quietly decided the eventual market structure.

The geography of the build-out will determine who captures the forecast's value. The United States leads in robot intelligence, foundation models, and the highest-profile startups, while China leads in bodies, components, and manufacturing throughput. If those strengths stay split, the likely outcome is an uneasy division where American software increasingly runs on hardware assembled abroad. The forecast quietly assumes a large, healthy market without resolving the more consequential question of which nation's companies actually book the revenue when the curve goes vertical.

Hidden Insight: The Forecast Is a Capital Magnet, Not a Crystal Ball

The temptation is to argue about whether 3 billion humanoids by 2060 is right. That misses the function of the number. Long-horizon forecasts from major banks are less prediction than coordination device: they give institutional investors permission and a framework to allocate, which pulls capital forward and accelerates the build-out the forecast describes. The precise 2060 figure will almost certainly be wrong; its job is not to be accurate but to be mobilizing, and on that measure it is already working.

The number that actually decides everything is not in the headline at all. It is the labor-cost crossover, the point at which a humanoid's all-in hourly cost drops below the human wage for a given task. Agility's Robots-as-a-Service pricing and Unitree's sub-$18,000 body are both attacks on that crossover from opposite ends, one through a service model, the other through hardware cost. The year humanoids become cheaper than people for warehouse work is the year the adoption curve goes vertical, and that year is a function of manufacturing and financing, not of any breakthrough in robot intelligence.

The bear case, however, is substantial, and the history of technology forecasting is littered with confident curves that flattened. Analysts predicted full self-driving cars by 2020 and a flying-car market that never arrived; the risk is that humanoids hit the same wall of edge cases, safety incidents, and unit economics that look great in a model and ugly on a factory floor. Skeptics point out that a 2060 forecast is conveniently unfalsifiable for decades, that early deployments remain measured in the low thousands, and that a single high-profile robot-related injury could freeze consumer adoption overnight.

There is a further underpriced risk in the supply story. If the 1.2-million-unit market materializes but is satisfied by Chinese manufacturers, Western investors buying into the bank's growth narrative may capture far less of the value than the forecast implies, while bearing the regulatory and security complications of foreign-made robots in sensitive settings. The optimistic top-line number can be directionally right while the distribution of who profits turns out very differently from what a US bank's clients are hoping to buy into.

Separating the two timescales the forecast blends together matters, because they demand different responses. The 2026-to-2030 ramp is concrete enough to underwrite investment and workforce planning today; the 2060 endpoint is a narrative anchor that mostly serves to make the near-term ramp feel inevitable. Treating them as one continuous certainty is how investors talk themselves into overpaying, and how policymakers talk themselves into doing nothing until the disruption is already at the door. The discipline is to act on the near curve while holding the far one loosely.

What to Watch Next

Over the next 90 days, watch whether other major banks and research houses publish competing humanoid models, since a chorus of institutional forecasts is what converts a single report into an investable consensus. Watch the early-2026 deployment counts against the 90,000-unit projection; running ahead or behind that pace is the first real-world check on the curve. And watch component pricing, because falling actuator and sensor costs are the mechanism that makes every later number on the chart possible.

Over the next one to two years, track the labor-cost crossover directly: any credible disclosure of a humanoid working a task below the local human wage is the single most important leading indicator in the entire sector. Watch for the first consumer-grade humanoid that does genuinely useful household work rather than demos, the event that would validate the home-ownership thesis underpinning the 2060 figure. And watch policy, since regulation, tariffs, and safety standards will shape how fast and from where the 1.2-million-unit market gets supplied.

The mental model to keep is that forecasts like this are best read as maps of capital intent, not prophecy. The useful signal is not the distant 3-billion-unit destination but the near-term ramp, the funding flows it unlocks, and the cost crossovers that will either validate or break the curve. Follow the unit economics and the shipment counts, treat the 2060 headline as marketing, and you will see the humanoid market for what it is becoming long before the prediction can be judged right or wrong. The eventual winners will show up in shipment data and falling unit costs years before any 2060 scoreboard is ever settled.

The 2060 headline is bait; the number that decides everything is the hour a robot becomes cheaper than a human.

---