Beijing now demands recurring revenue before a humanoid robot maker can list

Unitree took about 6.1 billion yuan off the market on August 19. At the first-day peak the market valued the company at roughly 445 billion yuan. The cash that reached the balance sheet was about 1.4 percent of the paper value attached to it. Twenty-one days later the China Securities Regulatory Commission published guidance telling future applicants in humanoid robotics to show recurring revenue, a credible path to profitability and genuine technological differentiation.

Crypto Briefing's own figures do not square. A fivefold rise from a roughly $9 billion debut valuation gives about $45 billion, not the $66 billion peak the same report cites, which implies a first-day move nearer 7.3 times. And 445 billion yuan down to 190 billion is a 57 percent decline, against the approximately 45 percent described. In dollars, the fall from peak is about $36 billion, against the roughly $30 billion of paper wealth the report says was wiped out. Retail investors who bought near the peak absorbed the bulk of those losses, according to Crypto Briefing.

Deep Robotics and Leju Robotics both have applications pending, and both are unprofitable. Under the old STAR Market framework that was not necessarily disqualifying, because the board was built to accommodate pre-profit tech companies. The report describes the new requirements in general terms and does not give numeric thresholds for recurring revenue or profitability, so how hard the gate bites is not yet knowable from the record.

What is knowable is the price the sector was carrying into it. Robotics companies were trading at multiples around 40 times forward earnings earlier in 2026. That is a forward earnings yield of 2.5 percent, and a buyer paying that is paying for earnings that have not arrived. Somewhere between 100 and 150 companies sit in the Chinese humanoid sector, many of them living on research partnerships, education applications and pilot projects that have not scaled into repeatable revenue.

I'd expect the repricing to land on the last private round. The marginal pre-IPO buyer was underwriting a specific exit at a specific multiple, and that exit now needs several more quarters of trading history. Two ways that could be wrong. Robotics is a priority sector in China's 15th Five-Year Plan covering 2026 through 2030, so a regulator gating on revenue may be sequencing listings and not closing them. That is close to Crypto Briefing's reading, which is that firms which would have listed in late 2026 or early 2027 will need extra quarters of traction. The second is harder for the companies. A company cannot manufacture revenue traction because a regulator asked for it, and over 100 firms are competing for the same limited pool of pilot projects and government contracts.

The claim that tighter criteria remove the exit ramp financing these companies is Crypto Briefing's analysis, not a CSRC statement.