Global humanoid robot shipments jumped 272% year-over-year in the first half of 2026, reaching 19,100 units, according to a new report from Smart Analytics Global. That is nearly a quadrupling in twelve months, and the firm expects the full year to hit near 60,000 units, on the way to a projected half a million by 2030.
Last year, humanoid robot makers shipped just 13,317 robots, heavily weighted to the latter half of the year. We’ve already surpassed those numbers today.
Perhaps more startling, however, are these two numbers: 97% and 85%. Chinese vendors accounted for more than 97% of all humanoid robots shipped worldwide in the first half of the year, and Chinese buyers absorbed more than 85% of global demand.
A few key observations: everyone grew as a rising tide lifted all boats. Also, however, as the shipment numbers grow, the American humanoid robot vendors that showed up in last year’s report are now all grouped into the “Others” category. That may not be as catastrophic as it sounds: more on that later.
A changing of the guard at the top
The most striking shift inside the numbers is who’s now in first place.
AgiBot, the Shanghai-based startup founded by a former Huawei prodigy, shipped roughly 8,400 humanoid robots in the first half — good for 44% of the global market and a staggering 562% jump from a year earlier. That vaulted it past Unitree, the reigning champion, which shipped about 5,900 units for a 31% share.
Together those two companies now account for three-quarters of every humanoid robot on Earth.
“Global humanoid robot shipments nearly quadrupled in 1H 2026, but an equally important shift is taking place in how these robots are being deployed,” said Linda Sui, founder and principal at Smart Analytics Global, in the report.
AgiBot’s edge came from breadth.
Where Unitree leaned heavily on its popular G1 — a compact, affordable half-size humanoid that has become the default machine for university labs, research groups and viral dancing-robot videos — AgiBot fielded a full lineup: full-size bipedal A-series robots, smaller X-series units, and wheeled G-series models aimed at factories and warehouses.
It has also begun pushing into Europe, with early deployments in the U.K. and Germany. (A rumored planned push into the United States is likely on hold right now, thanks to the FCC foreign robot ban.)
The competitive stakes are about to get higher. Both companies are racing toward public listings, and Unitree has already set its Shanghai Star Market IPO at 150.8 yuan per share, aiming to raise roughly 6.1 billion yuan — about $904 million. Behind the leaders, a long tail of Chinese challengers is forming: Galbot shipped around 900 wheeled units, UBTech about 700, and Leju roughly 600, with Fourier and NOETIX also in the mix.
Robots are finally going to work
The other headline in the report is qualitative. These machines are increasingly doing jobs, not demos.
“Industrial and commercial applications accounted for more than 70% of shipments, up from approximately 50% a year earlier,” Sui noted. Robots are moving onto assembly lines, into logistics facilities and across commercial floors. That’s a key signal that the humanoid robot market is at least starting to mature from novelty – and dance shows – toward utility.
And revenue is following. Smart Analytics Global pegs 2026 humanoid revenue at around $1.6 billion, roughly doubling to $3 billion in 2027.
The wall going up in Washington
All of which lands at an interesting time for the United States.
In late July, the Federal Communications Commission moved to ban new imports of foreign-made humanoid and quadruped robots, citing cybersecurity and national security risks. The order names Unitree and AgiBot directly, and points to the risks of depending on machines packed with cameras, microphones and network connectivity that are built by a foreign power.
The ban applies to new versions of these products rather than devices already in use or previously approved, so it is more a wall against the future than a rollback of the present. But the intent is unmistakable: keep the next generation of Chinese robots out of American homes, businesses and government facilities.
The significance here, of course, is that if 97% of the world’s humanoid robots are Chinese-built, a ban on Chinese robots is functionally a ban on humanoid robots. American companies like Tesla, Figure, 1X, Agility Robotics and Boston Dynamics are building real technology, but none is shipping at anything close to Chinese volumes.
Why it’s too early to sound the alarm bell
I’m not personally ringing the alarm bells just yet for North American and European robot makers.
It’s worth remembering what these shipment numbers do and don’t measure. They count units out the door, not the quality of the technology behind them, and not who’s actually positioned to win the market that matters. I’m tracking over 400 humanoid robot makers globally and on the technology itself, my personal belief is that the American robots are right there.
Figure AI, Boston Dynamics, Agility, Apptronik, 1X: in capability, dexterity, and physical AI, U.S. machines are at least as good as what’s shipping from China, and in some cases better. Europe has companies in Humanoid and Neura Robotics that are right there as well.
The gap in the sales chart isn’t a gap in engineering. To a large extent, it’s a gap in strategy.
What China’s leaders are doing is shipping. What the American and European players are largely doing is building. And that’s a deliberate choice, not a failure.
They could be pushing robots out the door right now. Scale is a decision before it’s a market reality. They’re holding back until the product is genuinely reliable, genuinely good, and as close to guaranteed economically viable for customers. Scaling early is a massive risk: you push out too many robots with too poor capability, and you waste a lot of your capital and destroy your brand. Scaling too late is a massive risk too: competitors may eat up the market before you get there.
American humanoid robot makers are aiming for the Goldilocks solution: just the right time, when their robots do real jobs for real customers and the economics actually pencil out.
Going broad and chasing volume before you’ve hit that bar is how you flood the market with machines that disappoint, burn trust and generate return-to-senders instead of revenue.
There are real challenges ahead on manufacturing scale, supply chains, and cost. Those are not trivial, and China’s head start on deployment data compounds over time. But shipping robots that aren’t yet good enough is dangerous too.
The volume race is real, and it matters. But it just isn’t the whole race, and I’m considerably less worried than the raw numbers suggest. If nothing else, the FCC foreign robot ban gives domestic producers even more time to respond.