The Trump administration is escalating its effort to keep Chinese technology out of sensitive US environments, moving to block new foreign-produced humanoid robots and connected mobile machines from entering the country.
The policy formally applies to overseas-made equipment rather than China alone – but its political and commercial target is clear: China’s fast-growing robotics industry.
The US communications regulator has added new foreign-produced advanced robots and connected power inverters to a government list of equipment considered an unacceptable national-security risk.
In practice, affected products will generally be unable to gain the approval needed to be imported, marketed or sold in the US.
The move applies to future models, rather than equipment already authorised for sale.
Posting on LinkedIn, US-based trade association and advocacy group the Association for Uncrewed Vehicle Systems International (AUVSI) welcomed the decision.
“The AUVSI welcomes strong, risk-based action against the national security risks of [China’s] robotics playbook.
“The threat is real and documented. While these restrictions are necessary, alone they are not sufficient. AUVSI calls on Congress to codify today’s action through the GUARD Act and to pair it with real investment in American robotics capacity.”
Why Connected Robots Have Become A Security Issue
Humanoid robots are frequently presented as the physical face of AI: machines able to move through warehouses, factories, retail sites and public spaces while carrying out tasks designed for people.
However, the same features that make them attractive to employers raise difficult security questions.
The US government’s definition of the affected devices is broad. It covers mobile robots, including humanoids and animal-like quadrupeds, capable of navigating autonomously or responding to sensor data. The machines operate away from a human operator and use sensors, network connections and software enabling autonomous movement, data collection or remote command and control.
That moves the debate beyond whether a robot looks human. A connected machine can carry cameras, microphones, environmental sensors, mapping tools and cloud links through an organisation’s operational environment. It may also receive remote support and software updates.
US authorities say these devices could be exploited to surveil Americans, support foreign intelligence services or remotely commandeer machines.
In a workplace, the concern is not merely one faulty robot. It is the possibility that a compromised connected fleet could collect sensitive information, map a secure facility, interrupt physical operations or provide a route into the wider corporate network.
This sits within a larger US security concern around Chinese technology in critical infrastructure.
The Volt Typhoon campaign, disclosed in 2023, involved alleged Chinese state-linked efforts to gain persistent access to networks supporting US critical infrastructure through compromised internet-connected devices.
The Rules Are Broader Than China
The restriction applies to new foreign-produced advanced robotic devices and power inverters. A manufacturer’s nationality is not, by itself, the deciding factor. The regulator says it will assess whether a product is foreign-produced under US procurement definitions.
That means a robot offered by a US-headquartered business could potentially be caught if its production arrangements do not meet the domestic threshold.
At the same time, a conditional approval route is available where authorities determine that a specific device, or class of devices, does not present an unacceptable risk.
Some reports have suggested that many non-Chinese suppliers could receive exemptions, potentially narrowing the rule’s real-world effect.
China’s developers have moved quickly in the humanoid market, supported by deep manufacturing supply chains, component availability and a vast domestic market. Companies including Unitree and AgiBot have become prominent names as the industry moves from laboratory demonstrations towards early commercial applications.
US Companies Are Testing Humanoids At Work
Looking at the bigger picture, humanoids still remain at an early stage in the US market.
Most activity is concentrated in pilots and closely controlled deployments, rather than fleets operating independently across factories or fulfilment centres.
The use cases are usually narrow and repetitive: moving materials, handling containers, inserting components, inspecting parts or supporting workers with physical tasks in sites built for people.
BMW has tested California-based Figure AI’s Figure 02 humanoid robots at its Spartanburg, South Carolina plant.
The company said an 11-month deployment saw robots insert sheet-metal components into fixtures for chassis assembly and supported production of more than 30,000 BMW X3 vehicles.
Amazon also uses Agility Robotics’ Digit in fulfilment-centre workflows, assigning it the task of moving empty tote boxes after goods have been removed.
What Enterprise Buyers Need To Consider
For organisations that already operate authorised foreign-made robots, the immediate impact should be limited.
Existing models can continue to be imported, marketed, sold and used, while basic software and firmware updates remain permitted.
The greater uncertainty lies in future procurement.
Businesses considering humanoids, autonomous mobile robots or quadrupeds should check whether a specific model has the required US approval, whether planned product updates may constitute a new model, and where the robot and its critical components are made.
They should also examine the security model behind the hardware.
Questions include where sensor data is stored, whether remote vendor access can be restricted, who controls software updates, how the robot is segmented from core business systems and what continuity plans exist if a supplier loses US market access.
Vendors can seek conditional approval where authorities find a product does not pose an unacceptable risk. But that process introduces another regulatory variable for buyers and suppliers operating in a market that is already moving quickly.
China has condemned the decision and urged the US to stop what it characterises as the smearing of Chinese companies. For enterprise technology leaders, though, the immediate lesson is straightforward: robotics procurement is becoming as much a question of supply-chain security and geopolitics as it is of capability, cost and return on investment.
