A robotics market emerges when three factors come together: capital to finance development; market access that allows for volume production; and technical capability that justifies its use. For half a century, these three factors grew more or less in step in the West—German and Japanese manufacturers developed, sold, and improved their products in the same markets where they were based.

August clearly demonstrated that this synchronized pace has fallen out of step. Capital is concentrating in Asia, the most important Western sales market is drawing a line, and the leap in capability is being showcased at Chinese trade shows, not European ones. Three events within four weeks each mark one of these trends. Taken individually, they are news items. Taken together, they describe a shift that directly affects European procurers—because the question is no longer when humanoids will be ready for deployment, but under whose terms they will be.

Shanghai: The Price

On August 19, Unitree Robotics made its debut on the Shanghai STAR Market—marking the first time a humanoid robot manufacturer has listed on a stock exchange in mainland China. The stock opened at 1,100 yuan—629 percent above the offering price of 150.80 yuan—before falling back to close at 845 yuan, representing a gain of 460 percent and a market capitalization of approximately 342 billion yuan. Ten percent of the expanded capital was sold, raising approximately 6.1 billion yuan.

The fundamentals don't support that. Revenue climbed to 1.70 billion yuan in 2025 from 392.77 million the previous year; net income was 278.21 million yuan; and over 5,000 humanoid units were shipped. Reuters Breakingviews calculated that the closing price was about 857 times the earnings expected for 2026; the average on the STAR Market is around 130. The company is profitable—but its machines are primarily used in research and education, not in commercial applications. The price clearly reflected scarcity, not earnings power.

Washington: the Border

Three weeks earlier, U.S. regulators had closed a loophole. On July 28, the Federal Communications Commission added “advanced robotic devices”—defined as mobile robots such as humanoids and quadrupeds—as well as foreign-manufactured networked inverters to its Covered List. As a result, new models will not receive FCC approval and may not be imported, marketed, or sold in the U.S. Crucially for European manufacturers: the measure is not aimed solely at China. “Foreign-produced” refers to anything that is not a “domestic end product” under U.S. procurement law. Models approved before July 28 remain unaffected.

As things stand today, traditional industrial robot arms are not included in this category. However, no one should count on this: This was the third category-based listing within eight months, following drones and consumer routers. The IFR has been officially monitoring the situation since August 7.

Beijing: The Capabilities

Coinciding with the IPO, the World Robot Conference opened in Beijing. Over 300 companies—69 percent more than the previous year—showcased more than 3,000 exhibits, including over 300 premieres. The turning point was not the number of humanoids, but the fact that they were working on-site: handling circuit boards, sorting logistics, and loading and unloading auto parts.

Unitree founder Wang Xingxing himself set the benchmark. The mass market will begin when robots can complete 80 percent of the tasks assigned to them via voice command in unfamiliar environments—a process that will take two to ten years. A timeframe that leaves everything open.

What Europe has left is its lead in implementation

Of the three key factors, Europe has none in excess: neither the valuation premium nor the protective barrier. What remains is its lead in implementation—decades of experience in integrating automation into ongoing production. It is real, and it is dwindling.

In practical terms, this means three things. First, the question of origin must be part of every supplier review—at the level of the FCC ID and approval date, not at the level of the company’s headquarters. Second, no trade show video can replace a reference installation with shift logs. Third, a stock market valuation measures expectations, not maturity. Anyone who confuses the two is buying a narrative instead of a system.