Invest1 distinct publisher3 min readPublished
A Chinese humanoid maker popped 460% on debut the same week Evergrande's founder was jailed for life. Solar, batteries and cars already show what state-funded buildouts do to margins.
The Investor · Invest desk

Compiled by The InvestorSomething wrong?How this is made
Solar has already run this experiment. Chinese module prices fell by half in 2023 and another 25% in 2024 [8], which compounds to roughly 62.5% below where they started [9]. That happened in a market where Chinese producers hold more than 90% of global supply, and their profits were eviscerated anyway [7]. Module demand was real and growing the whole time. The margin went regardless, because the capital funding the capacity was not priced on returns.
Batteries repeated it: global capacity, China-dominated, reached almost three times actual demand [10]. In cars, Brad Setser of the Council on Foreign Relations, previously at the US Treasury, puts Chinese capacity at 55 million units, close to two-thirds of global demand and rising quickly [11].
Each of those buildouts at least had a product with paying customers. Humanoids do not yet. According to Semafor, nobody really knows whether there is a commercial application for the machines [16], and what has captivated the Chinese public is spectacle: Unitree's "Superman" model out-accelerates Usain Bolt and leaps six feet six inches [17]. Behind it sit more than 150 companies making similar products, 50 of which have listed or plan to [12], one in three of the field [14]. A Beijing venture capital executive told the Financial Times that humanoid robotics is approaching the peak of the hype cycle [15]. For the pre-IPO holders sitting on windfall gains [16], the exit is the next listing, and the next listing is bidding for the same buyers as the other 49.
The political leg works the same way. Semafor's reading is that Hui Ka Yan, whose Evergrande collapsed under more than $300 billion of liabilities [4], was not a deviation from the development model but its instrument: real estate peaked at 25% of the Chinese economy [5] and left up to 90 million vacant apartments [6], built by developers who reasonably assumed Beijing would not spoil its own party. Desmond Shum, author of Red Roulette, posted on X that "the system created the incentives, rewarded the behavior, and profited from the boom" [19]. The bill for that arrangement has landed personally, and repeatedly. GOME's Huang Guangyu was jailed for bribery and insider trading after the 2000s retail boom [20]. Wu Xiaohui, owner of New York's Waldorf Astoria, was jailed when the outbound investment wave stopped [21]. Jack Ma escaped with a large fine on Alibaba and the loss of perhaps half his assets [22].
Wang Xingxing's roughly $16 billion is paper [2], and Semafor's point is that private Chinese fortunes of this kind are precarious [24]. The more useful number for outside money is the entry price: a 460% first-day gain [1] means anyone buying at the close paid 5.6 times the offer [23]. Pre-IPO investors have already been paid for the overcapacity risk. Whoever bought at 5.6 times owns it.
Ranked by verification strength, evidence, and original report placement.
Unitree Robotics shares popped 460% in their market debut.
Solar module prices plunged by half in 2023 and 25% again in 2024.
More than 150 companies are making similar humanoid robot products, and 50 have either listed or plan to.
After the listing, Unitree founder Wang Xingxing is worth around $16 billion, at least on paper.
On successive days last week, investors hailed Unitree's debut and a court sentenced fallen real estate tycoon Hui Ka Yan to life in jail.
Evergrande, founded by Hui Ka Yan, collapsed under liabilities of more than $300 billion.
Follow any of these and your For You feed starts watching them — no settings page required.
Evidence-backed comparisons of source perspectives and observed adoption signals. Read the methodology
Which Builder, Operator, and Investor concerns the observed source mix emphasized—not a truth score.
Evidence, demonstrated adoption, hype gap, incentives, and confidence are assessed independently, each on its own current evidence. How these are measured.
Single signed column, mixed sourcing
All material comes from one named-opinion Semafor column. Some items carry real provenance: Brad Setser is named for the 55 million car capacity figure, Desmond Shum is quoted from X, and the hype-cycle quote is attributed to the Financial Times. But the headline market facts have no filing or exchange citation, the state-planner bubble warning names no agency or document, and no Unitree financials appear. There is no second publisher to corroborate anything.
Capital-market uptake, no product deployment
What is observable is financial adoption, not usage: one high-profile debut and roughly 50 of 150-plus humanoid makers listed or listing. On the product side the source offers only athletic demonstrations and an explicit statement that no commercial application is established, with no customers, pilots or shipment volumes. Adjacent sectors show demand shortfalls rather than uptake, with battery capacity near three times demand and solar prices collapsing.
Valuation far ahead of demonstrated commercial use
The underlying story shows a 460% debut, a $16 billion paper fortune and a listing wave set against zero disclosed revenue, customers or shipments, plus a stated absence of any known commercial application. Sector precedent in solar, batteries and autos shows this pattern ending in destroyed margins, and both an anonymous Beijing VC and state planners are cited warning of a peak or bubble. The column itself is skeptical rather than promotional, so the gap sits in the market's claims, not in the coverage; it is not scored higher because the crowding, price and capacity figures cited are specific and consistent.
Windfall holders, state promoters and a skeptical columnist
Named and unnamed actors in the story have strong stakes: Unitree pre-IPO investors sit on a windfall from the same debut whose durability is in question, 50 peer companies are seeking or holding listings, and state actors both directed the resources and now warn of a bubble. A quoted commentator is an author with a book on Chinese elite capitalism, and the anonymous VC's incentives cannot be checked. The publisher is a signed opinion column with a thesis to advance, though no financial interest is disclosed or evident.
Coherent argument, thin corroboration
The factual spine, market debut, Evergrande liabilities, sector price and capacity figures, and an analyst-attributed auto capacity number, is internally consistent and the arithmetic derivations check out. Confidence is nonetheless capped by single-publisher, single-source coverage, an opinion format, absence of any company financials, and two unverifiable attributions (state planners, anonymous VC) that carry much of the bubble claim.
product
Unitree's $9bn Shanghai listing prices the humanoid story, not the robot-dog business1 distinct publisher
invest
Unitree's $905M Shanghai listing prices humanoids at 35x sales while profit halves1 distinct publisher
invest
Google Ships Flash Instead of Pro While OpenAI Loses Its Two Best Operators1 distinct publisher
product
Nvidia's H200 finally lands in China at about 1% of the order book2 distinct publishers
Distinct publishers with included, body-backed reporting in this cluster.
1 article · August 25, 2026