Humanoid robot shipments to hit 300,000 to 400,000 in US by 2030: JPMorgan

The humanoid robot demand curve is ramping sharply

This is kind of a big deal: JPMorgan sees 300,000 to 400,000 humanoid robots rolling out of U.S. factories by 2030. That's hockey-stick style growth. By 2035, JPMorgan is forecasting we're at 1.5 million, and the component market alone is approaching $30 billion annually.

As reported by Reuters on November 12, the bank's analysts called humanoid robotics an industry that has "hit an inflection point." The underlying research note has not been published publicly, but the numbers Reuters attributed to it are specific enough to take seriously.

The gap between that vision and current reality is significant, of course. Most U.S. humanoid manufacturers outside Tesla target 10,000 to 30,000 units per year of production capacity. Actual annual output for many is under 100 units, and Figure AI has produced about 1,000 Figure 03 unites, total.

The industry is building toward those JPMorgan numbers, but it is not there yet, and JPMorgan acknowledges it.

The key finding, therefore, is critical: supply is the binding constraint, not demand.

The numbers and the caveats

Reuters flagged an important ambiguity: it could not confirm whether JPMorgan's 2030 and 2035 unit projections are annual output figures or cumulative totals. That distinction matters enormously, of course: 400,000 annual units by 2030 is a very different claim than 400,000 total units ever shipped by 2030.

I'm guessing it's cumulative, because annual would be pretty crazy.

What is clear from the other data points in the report: costs are falling at 40% per year, faster than the previously projected 15 to 20%. Agility Robotics' Digit started at approximately $150,000 bill of materials; JPMorgan sees that dropping toward $30,000 at 10,000 units per year. At that price point, with robots running multiple shifts and generating an estimated $100,000 in annual savings per customer, the return-on-investment math starts to work broadly across manufacturing and logistics.

The global context of course is that China is leading today.

Unitree and AgiBot accounted for roughly 80% of 2025 global volume at around 13,000 total units. A Leju Robotics factory in Guangdong reached a production rate of one humanoid robot every 30 minutes in late March 2026.

Supply is the constraint, not demand

JPMorgan identified four bottlenecks:

1. hands and grippers
2. actuators
3. AI training data and compute
4. raw production capacity
   

These are hardware and manufacturing constraints, not demand constraints. The report's implicit claim is that buyers are ready, the technology is approaching viability, and the limiting factor is the ability to produce robots fast enough.

That is (very) a bullish framing. It says the market's main problem is an addressable manufacturing scale problem rather than a fundamental technology or adoption problem. It implies that companies solving production capacity will capture outsize returns, and that the race to scale factories is as important as the race to improve AI.

Named US platforms in the report: Tesla Optimus, Agility Robotics Digit, Boston Dynamics Atlas, Figure AI, Apptronik Apollo. Component suppliers identified as likely beneficiaries: Timken, ROBOTIS, Leader Drive, Orbbec, and Cognex. JPMorgan separately initiated coverage on Agility Robotics (NASDAQ: AGLT) with a Neutral rating, which signals it sees near-term upside as already priced in.

Wall Street is paying attention, big time

Wall Street is tuning into humanoids.

• Goldman Sachs put the humanoid market at $38 billion by 2035 in a 2024 report.
• Morgan Stanley has published similar projections.
• Now JPMorgan is weighing in with numbers specific enough to generate significant media coverage.
  

When multiple major investment banks have assigned dedicated analyst resources to humanoid robotics, the category has crossed a threshold: this is no longer speculative enough to be classified as future tech. Institutions are building models on it.

The practical implication is that when Wall Street covers your sector this explicitly, the fundraising environment changes. Companies that can show production scale metrics, not just AI demo quality, will attract institutional capital at favorable valuations.

JPMorgan flagged one policy risk worth noting: the Trump administration's national security classification of imported robotics components could lead to tariffs, adding cost uncertainty to supply chains that already depend heavily on Chinese actuators, sensors, and motors. That is potentially a tailwind for domestic component producers, but a constraint for companies that currently source offshore.

The 7,000 to 400,000 trajectory in five years requires manufacturing capacity to grow roughly 50x. That is the core of what JPMorgan is predicting. The industry's challenge now is whether the factories can keep pace with the forecast.

Sources

• Reuters — JPMorgan humanoid robotics inflection point
• Ground News — JPMorgan humanoid report coverage