The humanoid patent map: China, China, China, but USA has quality
If you want to know who's really building the humanoid future and not just who's shipping the slickest demo, patents are one good tell. A new patent map of the industry shows that China has filed its way to a commanding lead in sheer numbers, while the United States is capturing significant value.
That's the picture from a LexisNexis Intellectual Property Solutions analysis that combed through more than 26,000 patent families across the three technology pillars that make a humanoid actually work:
- interaction systems
- morphology (the physical body — limbs, joints, bipedal movement)
- control and planning architectures
Together they're the closest thing we have to an X-ray of where humanoid R&D is really concentrated.
Where the patents live
Let's start with morphology, the most humanoid-specific bucket of the three. It's the category tied directly to human-like bodies and walking on two legs. Here China isn't just ahead; it's lapping the field.
By the end of 2025, Chinese inventors held 73% of the world's active morphology-related patent families. South Korea trails a distant second at 11%, Japan third at 7%, and the U.S. fourth at just 5%.
Raw count, by inventor location (morphology):
- China — 73%
- South Korea — 11%
- Japan — 7%
- United States — 5%
But the LexisNexis Patent Asset Index also scores a portfolio on how strong the patents are: how often other inventions cite and build on them, and how widely they're protected across markets. Weigh the field that way and China's share drops to 63%, while the U.S. goes from fourth to second, capturing 11% of the most valuable morphology patents.
Japan lands at 10%, South Korea at 7%.
Portfolio strength (morphology):
- China — 63%
- United States — 11%
- Japan — 10%
- South Korea — 7%
The headline for anyone tracking the industry: American inventors hold roughly one-twentieth of the morphology patents but about one-tenth of the value, at least in this one key area for humanoids.
Across all three technology categories, U.S. patents posted the highest average quality per patent. That's fewer swings, but more home runs.
That said, when you have enough quantity, you generally win in quality too, and the Chinese companies are doing exactly that.
China's lead is policy, not luck
This isn't a one-year blip.
Over the past decade, Chinese patent activity roughly doubled in interaction systems, grew about 2.5x in morphology, and jumped roughly 5x in control and planning architectures. Morphology filings in particular were flat for years, then spiked in 2023 ... right as humanoids flipped from lab novelty to the hottest ticket in venture capital.
It tracks with Beijing's playbook.
China named robotics a strategic industry in its 14th Five-Year Plan back in 2021, singling out the exact subsystems humanoids depend on: integrated joints, electronic skin, bionic perception, intelligent control. Its newest Five-Year Plan, out in March, elevates embodied AI to one of six "future industries" the state intends to cultivate through 2030.
So while the quality-versus-quantity story is real, let's be honest: on both raw volume and total portfolio strength, China leads, and not by a hair.
Interesting: the strongest humanoid portfolios aren't the humanoid companies
Here's a twist:
Nine of the 10 organizations with the strongest patent positions across the three core technologies aren't humanoid specialists at all ... they're companies whose business goes well beyond two-legged robots. Often they're companies that may not even offer a humanoid robot.
Which makes sense once you accept that a humanoid isn't one invention. It's a stack: AI models, perception systems, batteries, compact actuators, safety systems, fleet software.
The strongest patents tend to sit in those enabling layers. The top ranks include industrial-robotics veterans Fanuc and Kawasaki Heavy Industries, AI and compute players Alphabet and Nvidia, surgical-robotics specialist Intuitive Surgical, Samsung and the Chinese Academy of Sciences.
For most of them, humanoid-specific patents are less than 10% of the total portfolio.
Zoom into dedicated humanoid startups ranked by patent strength and China's strength shows: six of the top 10 are Chinese, led by Fourier, Agibot and LimX Dynamics. The West's entries — Germany's Agile Robots and U.S. names Figure AI, Apptronik and Agility Robotics — make the top 10 list too.
The lone pure-play humanoid robot company near the summit of the overall rankings is China's UBTech.
Now factor in the U.S. import ban ...
Patents are a leading indicator: they show where the R&D is before it reaches a warehouse floor.
And the market on the other end is enormous — Goldman Sachs pegs a roughly $38 billion humanoid market by 2035, while Morgan Stanley floats $5 trillion by 2050 once supply chains and support are baked in.
Except the U.S. just moved to bar Chinese — and all foreign — humanoids and robot dogs from the American market on national-security grounds. That reshapes the commercial math, but it's worth being precise about what it does and doesn't touch.
A market-access ban and a patent lead are different animals.
Patents are territorial and indifferent to customs rules — China's 73% share of morphology filings sits in patent offices worldwide whether or not another Unitree unit ever clears U.S. ports. Cutting China off from its biggest export market squeezes commercialization, not invention. It could eventually dent innovation, but only if China's domestic and rest-of-world sales can't carry the commercialization load,
And that's unlikely in the near term.
For U.S. builders, the flip side is a gift: locking out Chinese hardware hands Figure, Apptronik and Agility a protected runway and a strong reason to develop the actuators, sensors and control stacks in-house. Over time that should boost U.S. filing volume, not just quality. The catch: an import ban doesn't stop Chinese firms from holding U.S. patents, so American manufacturers scaling into the gap may still have to design around — or license — Chinese IP on core morphology.
The most likely outcome is a race that splits down the middle: China compounding on volume, the U.S. leaning on quality and its lock on the AI-and-compute layers.
But the U.S. is about to start manufacturing serious quantity of its own, and China already has plenty of quality.
All we can say for sure is that it's going to get interesting.
Written by
I tell stories with data and build startups. I've raised angel and VC funds, led software companies and startups, and chronicled the rise of the mobile economy. Currently I write for Forbes, host several podcasts, and am starting up HumanoidMegaHub. I was TUNE's mobile economist and Singular's VP of Insights, built the research team at VentureBeat, and managed teams creating software for partners like Intel and Disney. In past lives I built websites, e-learning platforms, and mobile apps, and, a long time ago, was named to Folio's top 100 of the media industry's "most innovative entrepreneurs and market shaker-uppers.”