J.P. Morgan Humanoid Forecast: 400,000 US Units by 2030

By RWT Robotics Editorial Team Published November 12, 2026

CONFIRMED

J.P. Morgan projects that annual US humanoid robot production could reach 300,000 to 400,000 units by 2030 and about 1.5 million by 2035, Reuters reports in coverage syndicated by Investing.com. The bank also expects the market for key humanoid components to approach $30 billion by 2035. These are analyst projections about future production. They do not count robots that have been built or shipped.

2030 forecastCONFIRMED B J.P. Morgan projects 300,000 to 400,000 US units a year, per Reuters

2035 outlookCONFIRMED B about 1.5 million units and a component market near $30B, per Reuters

Maker targetsCONFIRMED B most US makers other than Tesla aim for 10,000 to 30,000 units a year, per Reuters

Actual US shipmentsNOT ANNOUNCED B the Reuters coverage gives bank forecasts, not delivery counts

What we know

Reuters reports that J.P. Morgan sees annual US humanoid production reaching 300,000 to 400,000 units by 2030. The 400,000 figure in our headline is the top of that range. The bank frames the growth as the industry moving from demonstrations to commercial deployment, and it projects about 1.5 million units a year by 2035.

The bank also looks at the supply chain. Reuters reports that J.P. Morgan expects the market for key components, specifically actuators, motors and vision systems, to approach $30 billion by 2035. Some summaries of the story give “near $30B” with no date attached. The Reuters-syndicated coverage dates it to 2035, so that is the figure we use. The bank favors suppliers including Timken, Robotis, Leader Drive, Orbbec and Cognex.

According to Reuters, J.P. Morgan says most US humanoid makers other than Tesla are targeting 10,000 to 30,000 units a year. The bank also says humanoid costs are nearing parity with labor costs. It estimates that multi-shift deployment could save a customer about $100,000 a year. It lists three main constraints: hardware, AI training data and compute, and production capacity.

Earlier J.P. Morgan research adds some context. It put unfilled US manufacturing jobs at about 462,000 today, rising to 1.6 million by 2030, and estimated a humanoid selling price of roughly $120,000. Separately, Morgan Stanley has forecast that China’s humanoid shipments will reach about 50,000 units this year.

Reuters also notes that the Trump administration has found that imported robotics and industrial machinery pose national security risks. That finding could lead to tariffs or other measures meant to boost domestic production.

What we don’t know yet

The most important missing number is today’s US baseline. The coverage gives targets and projections. It gives no count of humanoids currently built, shipped or working in US facilities. Without that, we cannot tell how steep the climb to 300,000 units really is. It could be an aggressive ramp or an enormous one.

We also don’t know Tesla’s role in the forecast. The bank sets Tesla apart from the 10,000 to 30,000 unit targets of other makers, but the coverage does not say how much of the 2030 range it assigns to Tesla.

The $100,000 savings figure needs care. It is described as savings per customer from multi-shift deployment, not per robot. The coverage does not say how many robots, shifts or hours that assumes, so it cannot be turned directly into a payback period against the earlier $120,000 price estimate.

Finally, “nearing parity with labor costs” is the bank’s characterization. The coverage does not show the underlying cost model, uptime assumptions or maintenance figures. Those details decide whether a robot actually pays for itself on a factory floor.

What it means

Start with the arithmetic, because the bank’s own numbers allow it. If most non-Tesla US makers top out at 30,000 units a year, reaching 400,000 would take more than a dozen of them running at the top of their targets. The other option is for Tesla to carry a very large share on its own. Neither is impossible by 2030. Both require manufacturing scale that humanoid companies have not yet shown in public.

The China comparison cuts in two directions. Morgan Stanley forecasts China’s humanoid shipments at about 50,000 units this year. That suggests volume humanoid production is already underway. It also means the US forecast for 2030 is roughly six to eight times that figure for all of China today. And actuators are one of the components J.P. Morgan expects to drive the $30 billion market.

That is where the tariff question matters. If Washington acts on its national security finding about imported robotics and machinery, US makers could face higher component costs or be pushed toward domestic suppliers that are not yet at scale. The constraints J.P. Morgan itself names, hardware and production capacity, sit right in that gap.

The labor argument is the strongest part of the bank’s case. A projected 1.6 million unfilled manufacturing jobs by 2030, per J.P. Morgan’s earlier research, is the kind of demand that justifies building factories. Demand alone does not ship robots, though. The bank’s third constraint, AI training data and compute, is the one that decides whether a humanoid can do useful work across a full shift without a person stepping in. Production volume means little if the robots are not doing productive work.

For readers tracking the field, the useful approach is to treat 400,000 as a ceiling scenario from one bank. It is not a trajectory the industry has demonstrated. The forecast is credible enough to watch because Reuters reports it from a major institution with stated assumptions. It becomes more than a forecast only when companies publish delivery counts and customers report hours worked. The figures that matter next are actual shipments, Tesla’s share, actuator sourcing, and any tariff decision. Until those arrive, the gap between today and 2030 is a projection, not a production line.

Follow every development in one place on our humanoid robot tracker.