A 58% haircut is not a rounding error. That is the gap IDC sees opening between the humanoid robotics revenue the U.S. market could generate by 2030 and what it will actually produce if Washington's restrictions on foreign robots and robot components stay in place. The projection, which circulated this week through market-data aggregators, frames the curbs less as a tariff story than as a market-structure story. American companies are being separated from the one country that currently builds humanoids in volume, buys them in volume, and manufactures the parts inside them at prices Western suppliers have struggled to match.
The numbers behind that split are easy to sketch even without IDC's full dataset. China's humanoid ecosystem, anchored by firms such as Unitree, UBTech, AgiBot, and Fourier, has spent the last two years pushing sub-$20,000 price points and shipping units by the thousands, while most U.S. developers remain in pilot-stage deployments with a few dozen robots at a time. Unitree's G1 became an informal reference platform for university labs worldwide precisely because it undercut Western alternatives by an order of magnitude. Restrictions that block imports of such machines, and the subsystems that go into them, remove a price anchor from the U.S. market. Higher prices mean slower adoption. Slower adoption means smaller revenue.
The harder problem sits below the robot level. A humanoid is largely a stack of motors, planetary and harmonic reducers, frameless torque actuators, rare-earth magnets, encoders, batteries, and sensors, and a startling share of that bill of materials traces back to Chinese suppliers. American humanoid developers, including Tesla with Optimus, Figure AI, Agility Robotics, and Apptronik, have all talked about vertical integration and domestic sourcing, but building a competitive actuator supply chain in the U.S. is a multi-year, capital-intensive undertaking. Rare-earth magnet processing alone remains overwhelmingly concentrated in China. If component access tightens while domestic capacity is still being built, the cost curve that humanoid makers promised investors, steep declines as volume ramps, could flatten before it ever bends downward.
The damage runs both directions, and that matters for how the two blocs evolve. Chinese humanoid makers lose the U.S. as an export market and, perhaps more significantly, as a source of high-end compute, with NVIDIA's Jetson Thor-class modules and the broader CUDA software ecosystem subject to their own export controls. That pushes Chinese developers toward domestic silicon and domestic foundation models, which in turn produces robots with incompatible software stacks, different safety certifications, and separate developer communities. For multinational customers, particularly automakers and logistics operators with plants on both sides of the Pacific, the prospect of qualifying two entirely different humanoid fleets is an operational headache that no one has budgeted for. The industry is drifting toward the same bifurcation that already defines telecom equipment and, increasingly, semiconductors.
Here is the observation that gets lost in the headline figure: a smaller U.S. market does not necessarily mean a weaker set of U.S. companies. Protected markets can be kind to incumbents. If foreign competition is shut out, American humanoid developers face less price pressure and potentially more room to command premiums, particularly in defense-adjacent, government, and critical-infrastructure applications where origin restrictions are a feature, not a bug. The risk is the opposite one. Without a low-cost foreign benchmark forcing discipline, U.S. firms could settle into slower iteration cycles, while Chinese manufacturers, competing ferociously against each other at home and eyeing Europe, the Middle East, and Southeast Asia, drive costs down and capability up. The 58% figure measures revenue forgone in the U.S. It does not measure the lead that scale-driven learning curves might hand to the other side.
What to Watch
Track the next round of Commerce Department and USTR actions on robotics components over the coming 30 to 90 days, particularly any carve-outs for actuators, reducers, and rare-earth magnets, since those would soften IDC's downside scenario. Watch Tesla's Optimus production updates and Figure AI's and Agility Robotics' commercial deployment announcements for evidence that domestic supply chains can hold costs in check. On the other side of the divide, follow Unitree's progress toward its planned public listing and the volume numbers from AgiBot and UBTech, which will show how fast Chinese makers are scaling without U.S. access. Finally, monitor NVIDIA's Jetson Thor shipment commentary and any new export-control language, because compute access may decide how far the two ecosystems diverge.
