On Wednesday, Unitree Robotics became the first humanoid robot company to list on China's A-share market. The shares were priced at 150.80 yuan in an offering that raised about 6.1 billion yuan, roughly $905 million, for a 10 percent stake, and they closed the first day at 845 yuan, up 460 percent. Unitree's market value at the close was about $50 billion. The stock touched a gain of more than 600 percent before settling. Online demand ran to more than eight thousand times the shares available, and the retail allocation rate, 0.0181 percent, was the lowest ever on the STAR Market. By arithmetic on the company's disclosed figures, the shares closed at roughly 1,200 times 2025 profit. That last number is the one to stop on.

A multiple like that is not a price on a business. It is a price on a category, and the category's price is being set somewhere else. The reference asset for Unitree's valuation is not its income statement. It is Tesla's stock, and inside it, a set of promises Elon Musk has made about a robot called Optimus. The debut also closed a loop between two governments that had already moved on the same thin evidence: Washington barred new foreign-made humanoid robots from its market three weeks earlier, and Beijing had spent years steering the sector, from industrial guidance to strategic investors. Both are bets on a future neither side has demonstrated. Unitree's first-day price is the first pure-play price on that contested future, denominated in another company's promises.

A debut that outran its own financials

Unitree's own numbers are real and unusual. Revenue rose from 392.77 million yuan in 2024 to 1.70 billion yuan in 2025, up more than fourfold, with net profit of 278.21 million yuan and first-half 2026 revenue near 1.1 billion yuan. Unitree says it shipped more than 5,000 humanoid units in 2025, first globally by its own count with about a third of the market, and had delivered roughly 18,000 bipedal robots by the end of July. Its G1 humanoid sells for about $16,000, or 99,000 yuan, a price that reset the category's floor, while the full-size H1 launched around 650,000 yuan; industry tallies put the average Chinese humanoid price near 600,000 yuan in 2023 and near 166,000 yuan in 2025, down roughly 70 percent. Founder Wang Xingxing controlled about a third of the shares before the listing.

None of this supports a price of 1,200 times earnings. Even at the issue price, the company was valued at about 219 times 2025 profit and 36 times 2025 sales, and the first day multiplied those figures by five and a half. Earning it requires profit to compound at a rate the disclosures do not evidence, because the product market is still unquantified. The G1, by industry accounts, has become a default platform in university robotics laboratories. Lab sales are a real business. They are not the $50 billion business. The pop, taken alone, was not unusual: Chinese listings this year have averaged a 279 percent first-day gain, and memory chip maker CXMT rose 466 percent on its debut last month, per Reuters. What is unusual is what the pop priced. To see that, look at the other side of the category, in Austin.

The reference price is set in Tesla's stock

At Tesla's 2024 annual shareholder meeting, Musk said Optimus could make Tesla worth $25 trillion, roughly half the S&P 500's value at the time, and potentially more valuable than everything else Tesla does. He supplied the arithmetic beneath the claim: a production cost near $10,000 and a selling price near $20,000, and a robot in every household as the eventual market. He has called his own optimism about Optimus pathological, worth recalling whenever his projections feed a valuation. The record since then separates promise from product. On Tesla's July 22 earnings call, Musk called Optimus the hardest product to scale manufacturing in Tesla's history, citing roughly 10,000 unique parts and no existing supply chain, and warned that the ramp would be long and flat. Tesla has decommissioned the Model S and Model X lines at Fremont to install its first Optimus production line, with production slated for around the end of this year. The first robots will gather training data inside Tesla's factories, per coverage of the call; none is doing production work, and none produces revenue.

Connect the two companies and the debut's logic appears. Unitree's $50 billion market value is about 0.2 percent of Musk's $25 trillion figure. If the category develops the way the Tesla story assumes, then the company that already ships the cheapest humanoids, in the largest numbers, is the cheapest way to own that outcome without paying Tesla's full price. That is the bull case for the debut, and it has real force: Unitree sells working robots today, at $16,000, and its 2025 revenue grew fourfold. The bear case is its mirror image: a 1,200 times earnings multiple on a company whose largest export market has just gated its new models is not a discount to anything. The two securities have become one trade in two currencies. Good Optimus news lifts Unitree's shares, bad news drags them down, and neither reaction says much about Unitree's own books. Price discovery for the first humanoid robot stock is happening inside an American car company's share price.

Both governments moved before the products arrived

The policy backdrop makes the same point in two registers. On July 28, the FCC's Public Safety and Homeland Security Bureau added foreign-produced advanced robotic devices to its Covered List, the register of equipment barred from the authorization needed to enter the United States market. The listing followed national security determinations by a White House-convened interagency body that the devices pose unacceptable risks to Americans. The rule covers new models only, and its reach is broad: anything heavier than 4.4 pounds that can navigate, sense its surroundings, and connect to a network, including robot vacuums. Exemptions can be granted only by the Pentagon, through January 2028. Chairman Brendan Carr said the agency was acting "in lock step with our national security agencies." The determinations cited documented incidents, including a September 2025 exploit that affected fleets of Unitree robots and raised the prospect of a self-propagating humanoid botnet. Reuters has reported the agency is expected to exempt most non-Chinese manufacturers. The United States was the largest export destination for Chinese-made humanoid robots last year, and Barclays has estimated Chinese manufacturers at roughly 85 percent of global humanoid shipments.

Beijing read the ban as it reads most export controls: as security stretched too far. The foreign ministry said it firmly opposes the overgeneralization of the national security concept and argued that protectionism will not improve American competitiveness; the commerce ministry threatened countermeasures. Three weeks later came the other half of the symmetry. China's Ministry of Industry and Information Technology issued its guidance on humanoid robots in late 2023, calling them a potentially disruptive product class following computers, smartphones, and electric vehicles, with targets for batch production by 2025 and for two or three globally influential enterprises; the sector sits inside the doctrine of "new quality productive forces." State media has projected a domestic humanoid market of 870 billion yuan by 2030, a projection without a track record. The listing itself ran on state machinery: strategic investors included the National Council for Social Security Fund, the AI lab DeepSeek, and state-linked firms such as CNPC Kunlun Capital, China Southern Power Grid, China Telecom Capital, and Tencent. Unitree went public on the opening day of the World Robot Conference in Beijing, where more than 150 new humanoid models are being unveiled.

None of this is sinister, and none of it is neutral. The listing was a way for Beijing's industrial policy to raise capital in public markets; the ban was a way for Washington's security policy to keep foreign hardware out. Both governments acted on the same thin record. There is no verified market for humanoid robots at scale, no verified answer to who would buy a million a year, for what work, at what maintenance cost. This analysis takes no position on whether the FCC's security assessment is correct or whether Beijing's industrial strategy is warranted; those are contested political questions. What is observable is that both sides treated the category as decisive before the category produced evidence.

The record is thin in both directions

The evidence that does exist cuts both ways. For Unitree: revenue up more than fourfold in a year, roughly 18,000 robots delivered, the lowest prices in the category, profitability that the previous "first humanoid robot stock" never reached, and about 44 percent of 2025 revenue from outside China. The company's updated prospectus noted that its current models already hold FCC certification and can keep selling in the United States, which took about 13 percent of 2025 revenue, down from nearly 20 percent in 2024. It also warned that revoked certifications or a broader trade confrontation could end the company's high-speed overseas growth. For the skeptics: the label has been worn before. UBTech listed in Hong Kong in December 2023, peaked near HK$328, and by early 2025 traded below its HK$90 issue price after lock-ups expired and institutional holders, including Tencent, reduced stakes. The label did not protect that stock. Unitree's defenders would answer that UBTech was loss-making and never had this growth or price point. Both statements can be true at once.

What neither company discloses is the quality of the demand: no verified public figure exists for how many humanoid robots are doing paid work, in either fleet, at what margin. Tesla's Optimus has no revenue; Unitree's sales are real but concentrated, by industry accounts, in laboratories, developers, and showcase demand. The multiples are being paid for a future that has been promised, subsidized, and banned into existence; a promise can be delayed, a subsidy withdrawn, a ban tightened. What Washington and Beijing agree on, without saying so, is that the category matters. That is why the FCC banned the hardware, why the state shepherded the listing, and why investors in Shanghai paid more than a thousand times earnings for a robot maker while investors in New York held Tesla partly on the same story.

One trade, two markets

The debut's meaning is now visible. The first pure-play price on the humanoid robot era was set in Shanghai, on a Chinese company, and the number owes more to the size of an American company's promise than to anything in the Chinese company's prospectus. The stock that trades at 1,200 times earnings and the stock that trades on a $25 trillion dream are the same position in two currencies, and the two governments that raced to respond, one by exclusion and one by encouragement, are betting the same unproven future in two different registers. Whatever you believe about the politics of robot competition, both sides would be served by the same thing: honest data on what these machines actually do, for whom, and at what cost. Until that exists, every price in the category is a conviction price, and conviction is not yet evidence. Unitree's first day was a triumph for the company and for the story. Whether it was a sound price will be decided by shipments, not headlines.

Primary sources

  1. Reuters coverage of the Unitree debut for the issue price, the 6.1 billion yuan raised, the first-day close, the market value, the oversubscription and record-low allocation rate, the 279 percent average first-day gain for Chinese listings this year, and the CXMT debut gain.
  2. Coverage of Unitree's updated prospectus for 2025 revenue, net profit, first-half 2026 revenue, shipment figures, the United States revenue share, the FCC certification status of existing models, and the escalation-risk warning.
  3. The FCC's Covered List action of July 28, 2026, and its legal analysis, including the national security determinations, the device definition, the Pentagon exemption window, the cited security incidents, and Chairman Brendan Carr's statement, and BBC reporting on Beijing's response to the ban.
  4. The Ministry of Industry and Information Technology's November 2023 humanoid robot guidance and state-media market projections, CNBC and Forbes coverage of Tesla's 2024 shareholder meeting for Musk's $25 trillion claim, and The Next Web's reporting on Tesla's July 22, 2026 earnings call for Musk's comments on the Optimus ramp.
  5. Reporting on UBTech's Hong Kong listing and its post-listing decline for the cautionary comparison.