A ban with an immediate trigger, and a slow-motion cost
On July 28, 2026, the FCC's Public Safety and Homeland Security Bureau added "foreign-produced advanced robotic devices" and "foreign-produced power inverters" to its Covered List, the roster of equipment barred from receiving the FCC authorization required to import, market, or sell a product in the United States. The public notice, DA 26-786, took effect immediately and implements a National Security Determination issued a day earlier by a White House-convened interagency body. Existing certified robots already in US homes and warehouses are unaffected; the ban hits new models seeking market entry.
The security rationale is not manufactured. The determination cites a vulnerability discovered earlier in 2026 that let remote actors reach into thousands of consumer robots worldwide — live camera feeds, microphone audio, mapped floor plans of private homes — through cloud-connected firmware controlled by the manufacturer. A humanoid robot or quadruped operating inside a factory, hospital, or military-adjacent facility is a mobile sensor platform with persistent network access; if the vendor or its home government can reach that firmware, the device is a standing intelligence and sabotage risk. That is a materially different threat model than a mislabeled t-shirt or a subsidized solar panel, and it is the same logic that has underpinned telecom equipment bans on Huawei and ZTE since the 2019 Secure and Trusted Communications Networks Act, the statute the FCC is again using here.
What the rule actually requires
To avoid the ban, a robot must be built toward the Buy American Act's "domestic end product" standard: assembled in the US with domestic components exceeding 65% of value through 2028, rising to 75% in 2029. Manufacturers who can't yet clear that bar can seek a Conditional Approval from the Department of War or Department of Homeland Security by submitting a detailed, time-bound plan to build or expand US manufacturing — applications are due by January 1, 2028. Robots under 2kg or communicating below 200 kilobits per second are exempt, which carves out toys and basic sensors but leaves nearly every commercial humanoid or quadruped platform inside the rule.
The problem is that the domestic supply chain the rule assumes doesn't exist yet at any real scale. Chinese vendors accounted for more than 97% of all humanoid robots shipped worldwide in the first half of 2026, according to Forbes' analysis of industry shipment data, with Chinese buyers absorbing roughly 85% of global demand over the same period — the concentration the news hook references, though it describes demand more than production capacity. Two companies, Agibot and Unitree, together account for the bulk of that output. Actuators, precision gearboxes, and vision sensors built at that volume are, for now, overwhelmingly a Chinese-manufactured commodity.
Founders are already stuck between the rule and the supply chain
Rest of World's interviews with US robotics founders capture the gap directly. Anto Patrex of CosmicBrain AI put it bluntly: "Look at Apple, it's made in China. How do you expect even more complicated products like robots to be built in America?" Elizabeth Williams of Gemma said prototyping without Chinese manufacturing partners "would have been impossible to do... here." Some companies have reportedly had employees carry components back from China in luggage rather than wait on formal procurement. Michael Perry of Persona AI — who previously ran DJI's North American business and has publicly welcomed the broader push to onshore robotics — still frames the current rule as incomplete: "You need to provide the carrot as well as the stick," he argued, faulting Washington for restricting imports without simultaneously funding the domestic component base the rule requires.
Not every incumbent agrees the timing is premature. Boston Dynamics' VP of policy, Brendan Schulman, endorsed the ban on LinkedIn as the opening move in a policy arc that will define the industry "for decades to come" — a defensible position for a company with an established US manufacturing and assembly footprint that smaller entrants lack.
The steelman, and where it breaks down
The case for urgency is genuine: robotics is following the same trajectory as telecom and solar, where years of import dependence made a later course correction far more disruptive than an earlier one would have been. Waiting for a domestic supply chain to mature organically, while Chinese-made humanoids get installed across US factories and warehouses, risks locking in exactly the dependency the rule is designed to prevent — and doing so with hardware that has cameras, microphones, and network access built in.
But a 65% domestic-content floor imposed with an 18-month runway, in a market where the qualifying supplier base is 3% of global shipments, doesn't accelerate onshoring so much as it freezes market entry for the startups least able to absorb the cost of building or qualifying US suppliers from scratch. The FCC's own conditional-approval pathway implicitly concedes this — it exists precisely because compliance isn't achievable on day one. A firm-specific, time-bound waiver track for companies with a credible US manufacturing plan would meet the security objective — foreign-controlled firmware still can't reach the device — without handing incumbents with existing domestic footprints a multi-year moat simply by virtue of having gotten there first. Congress and the Commission should pair the restriction with direct investment in the component suppliers — actuators, gearboxes, precision sensors — the rule assumes into existence, or the policy will onshore market concentration before it onshores manufacturing.