Unitree falls five sessions straight — 46% off its first-day high

Unitree closed at 591.53 yuan on the Shanghai STAR Market on August 26, 2026, down 1.87% and below 600 yuan on a closing basis for the first time since listing. It was the fifth consecutive down session. The stock is 46.2% below the 1,100 yuan it touched intraday on its August 19 debut, and market capitalisation has fallen to about 239.3 billion yuan — roughly 205.6 billion yuan, or about 40 trillion won, below the first-day peak.

The two steepest days did most of the work: down 18.7% on August 20 and 10.31% on August 24. Reuters characterised the slide as feeding bubble concerns around Chinese technology and robotics stocks.

The price was set twice, and the second one was the market’s

The offering priced at 150.80 yuan a share, which we covered when it came out on August 6. That price already implied 219.23 times 2025 earnings against a sector average of 38.56 times, and a market capitalisation near 61 billion yuan.

On day one the market took it to 444.9 billion yuan at the intraday high — about 7.3 times the offer valuation — before closing at 845 yuan. Everything since has been the correction of that number rather than a repudiation of the offer price. At 591.53 yuan the stock is still comfortably above where the underwriters set it. What has been erased is the day-one premium, not the listing.

The fundamentals are the argument, in both directions

Unitree is not a loss-making story stock. It made 278 million yuan of net profit on 1.699 billion yuan of 2025 revenue, at a 60.13% gross margin on its core business. Very few companies in this sector can say that.

The first quarter of 2026 is where the case gets harder. Revenue still grew 68.49% year on year — a number most companies would take — but that is a deceleration, and net profit excluding non-recurring items fell 52.55% as R&D and selling costs rose. Growth slowing while spending rises is the specific combination that a 219× multiple cannot absorb.

Underneath that is the composition problem the whole category shares. Unitree has world-class hardware and the strongest brand in quadrupeds, but a large share of current revenue comes from four-legged robots and research and education units, not from humanoids doing work in factories and homes. The humanoid revenue that justifies the multiple is still ahead of the technology — reliability, price, model quality and actual task competence all have to arrive first.

What this prices, beyond one stock

Competition is intensifying on both sides. Domestically, UBTech, XPeng and BYD are all pushing into humanoids alongside a long tail of startups; XPeng’s robotics unit raised over $900 million this week at a $6.3 billion valuation. Abroad, Tesla is the obvious comparison, and tightening US restrictions on Chinese humanoids are a live risk for a company with meaningful overseas exposure.

The useful reading is not that Unitree failed. It raised about 5.9 billion yuan net and has that money for AI models, robot platforms, new products and production capacity regardless of where the shares trade this week. The useful reading is that public markets have now, for the first time, applied ordinary earnings arithmetic to a humanoid company — and the answer came back at roughly half of what private enthusiasm had assumed. Every pre-IPO humanoid valuation in China now has a listed comparable, and it moves daily.