Skip to main content
Live
Main content

FTC bans foreign humanoid robots, extending AI protectionism to robotics

The sweeping import ban covers humanoids, quadrupeds, and wheeled robots — hitting Unitree just as it targets a $6 billion IPO.

Jaeden Schafer
Editor in Chief · · 5 min read
FTC bans foreign humanoid robots, extending AI protectionism to robotics

The Federal Trade Commission last week issued a sweeping ban on foreign imports of advanced humanoid, quadruped, and wheeled robots, pulling a nascent industry into the same protectionist bracket as chips and AI models. The order cites two justifications: that foreign-made robots collect sensitive data in homes and facilities, and that US robotics companies need shielding from Chinese competition to build a domestic supply chain. It lands as Unitree, China's leading humanoid maker, prepares an IPO this week targeting a nearly $6 billion valuation.

There is no US robotics company at that scale. Figure's humanoids are not selling at volume. 1X has not started shipping robots to homes. Boston Dynamics remains the reference-class US player, but its hardware sells at price points measured in six figures rather than four. The economics are stark: a four-legged Unitree robot runs about $4,600, while a comparable Boston Dynamics unit lists near $278,000 — a roughly 60x gap.

The FTC's cybersecurity rationale rests on incidents like one cited in the ruling in which an attacker took control of 7,000 robot vacuum cleaners. That risk profile scales as humanoids move from viral demos into workplaces and homes, where they capture audio, video, and floorplan data continuously. The order carves out enough exceptions that its practical enforcement footprint is unclear, but its signal is not.

Key facts

  • 01The FTC banned foreign imports of advanced humanoid, quadruped, and wheeled robots, citing national security and supply-chain concerns.
  • 02A four-legged Unitree robot costs around $4,600 versus roughly $278,000 for a comparable Boston Dynamics unit.
  • 0390% of recent US university robotics research papers relied on Unitree robots, per the Association for Advancing Automation.
  • 04Unitree is targeting a nearly $6 billion valuation in an IPO this week, with no US robotics peer at comparable scale.
  • 05The administration is separately weighing a ban on open-source Chinese AI models that could cost US businesses $25 billion in annual savings.

Some US robotics executives welcomed the move. Gavin Kenneally, CEO of Ghost Robotics, which builds four-legged inspection robots, said the cybersecurity risks from foreign-made robots are real and that stronger requirements would benefit customers and the domestic industry. Ghost Robotics is among a small group of US firms whose business case improves materially if Chinese hardware is walled off from federal and enterprise buyers.

If today's announcement encourages stronger cybersecurity and a more level competitive environment, that's good for customers and good for the robotics industry.
Gavin Kenneally, CEO of Ghost Robotics

The rest of the sector is more exposed. US academic labs and startup research teams have built their pipelines on cheap Chinese hardware, buying fleets of Unitree units to train models on tasks from folding laundry to flipping waffles. An internal review by the Association for Advancing Automation, a US robotics trade group, found that 90% of recent US university robotics research papers relied on Unitree robots.

Aaron Prather, director of market intelligence at the group, said the ruling creates a challenge for US humanoid researchers and noted that Chinese models offer the best price-to-capability ratio available. If replacement hardware from US vendors costs 40x to 60x more, the practical effect is fewer robots in labs, smaller training fleets, and slower iteration — the opposite of what a domestic industrial policy aims to produce.

Chinese models offer the best price-to-capability ratio available.
Aaron Prather, Director of market intelligence, Association for Advancing Automation

The robotics ban fits a broader pattern. The Trump administration is reportedly weighing restrictions on open-source Chinese AI models that rival systems from OpenAI and Anthropic at a fraction of the cost, a move that would block US businesses from an estimated $25 billion in annual savings. Together, the two measures extend AI protectionism beyond frontier labs and into the tooling layer that startups, researchers, and enterprises actually use day to day.

The technical bar for humanoids remains low. Robots still stumble, mishandle objects, and lag behind human dexterity on basic manipulation. Google last week released an AI model designed to help humanoids learn new tasks faster, with tying a trash bag as one of its more notable demonstrated skills — a meaningful step given how difficult robot hand control has been, but not a sign of an industry ready to absorb a 60x cost increase on its underlying hardware.

Related · from this week
US drone and robot curbs collide with China's 86% humanoid share
Jaeden Schafer · 5 min read →

The counter-argument to the ban is that the US robotics sector isn't yet capable of filling the gap. Figure and 1X are not shipping at volume. Boston Dynamics sells at institutional price points. Ghost Robotics is defense- and inspection-focused rather than a general humanoid supplier. Cutting off Unitree without a domestic alternative at comparable cost pushes research budgets, not just procurement budgets, into a tighter box — and research is where the next generation of US robotics companies is supposed to come from.

The FTC ruling reframes robotics as an AI-industry problem rather than a hardware-trade problem, which matters for how the sector will be regulated, funded, and acquired going forward. Expect defense-adjacent US robotics firms to see valuation lifts and procurement tailwinds, and expect the humanoid research pipeline at US universities to slow visibly over the next 12 months unless carve-outs preserve academic access. The bet the administration is making is that a domestic supply chain will emerge faster than the research gap will widen. On current hardware economics, that bet has a long way to run.

ShareXLinkedInEmail
AI Box

Every AI model. One chat.

The latest models from ChatGPT, Claude, Gemini, Sora, ElevenLabs — 80+ models in a single chat. Compare answers side by side. Pick the best one every time.

  • ChatGPT, Claude, Gemini, Grok, DeepSeek — in one chat
  • Generate images & video with Sora, Veo, Ideogram
  • Compare any two models side by side
  • From $8.99/mo · 80+ models, all included
Try AI Boxaibox.ai
Trusted by 3,000+ teams
Got a tip?

Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.

Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.

AI Box Daily briefingFree · Daily · No fluff

Stay ahead of everyone in AI.

The tightly edited AI news email engineers, founders, and investors actually open. One email. Every weekday. Five minutes to finish.

Loved by 10,000+ AI professionals
Free forever. Unsubscribe with one click.

The briefing read inside teams at

Keep reading

More from Security

US drone and robot curbs collide with China's 86% humanoid share
Business

US drone and robot curbs collide with China's 86% humanoid share

Washington tightened restrictions on Chinese drones and advanced robots this summer, but five Chinese firms shipped 86% of humanoids in H1 2026.

Jaeden Schafer5 min read
Google logo
Models

Google DeepMind ships Gemini Robotics ER 2 as a video-native brain for robots

The new embodied reasoning model tracks task progress from live video, orchestrates other robots, and runs at 4x the speed of larger rivals.

Jaeden Schafer5 min read
How modern AI is pushing robots from single tasks to general-purpose work
Analysis

How modern AI is pushing robots from single tasks to general-purpose work

Boston Dynamics and Physical Intelligence say reinforcement learning and foundation models have redrawn what robot autonomy means in 2026.

Jaeden Schafer5 min read