When Unitree Technology rang the bell on August 19, shares opened at 1,100 yuan, surging 629% and pushing its market value to 444.9 billion yuan, prompting widespread cheers about the "dawn of humanoid robots." Yet just four trading sessions later, the stock had retreated 45.1% from its opening price, erasing roughly 200 billion yuan in market capitalization. Capital has cast its vote with real money on the future of embodied intelligence, but it must now confront an awkward reality — Unitree's revenue still largely depends on selling research and teaching equipment, its truly profitable commercialization scenarios have yet to be proven, and the industry's long-awaited "ChatGPT moment" has not yet arrived.
At the moment the bell rang, paper wealth reached its peak. Based on the opening price, founder Wang Xingxing's roughly 121.4 million shares translated into a personal stake exceeding 130 billion yuan, catapulting him to the title of richest "post-90s" entrepreneur. The wealth effect extended far beyond the founder. Some Unitree employees became early beneficiaries, with 45 core staff members joining the millionaire ranks through strategic placement alone. In the secondary market, investors who secured an allotment saw maximum paper gains of nearly 480,000 yuan, setting a new record for single-lot profits since the registration system was introduced on the A-share market.
No one expected the celebration to last less than a day. On the first day of trading, the stock fell from its 1,100 yuan peak to close at 845 yuan, trapping those who had chased the rally intraday. When trading opened on August 20, Unitree Technology gapped down 6.51%, then continued to slide, touching an intraday low of 685.01 yuan before closing down 18.70%. Investors who had been celebrating their allotments the day before suddenly watched their paper gains shrink at a visible pace. By August 24, the fourth trading session after listing, Unitree Technology's intraday decline widened to more than 10%, closing at 603.08 yuan, down 10.3%, with a single-day market value loss exceeding 20 billion yuan. Total market capitalization had shrunk to 243.9 billion yuan, down 45.1% from the opening price — nearly halved. In just four trading sessions, about 200 billion yuan in market value had evaporated. Investor forums were filled with lamentations; some reported losing nearly half their money, while others called it a trap.
Economist Fu Peng noted in an interview that Unitree's valuation before listing had already far exceeded general market consensus, and the post-listing price surge and volatility were extreme. The robotics industry remains in its very early stages, and despite the completed IPO, he believes it is fundamentally a venture capital play of the primary market. Under the fervent expectations of the technology wave, it does not yet qualify as a long-term value or growth investment. "Ordinary investors can bet on the future with a small amount of capital, but if they seek stable investment returns, it is not a suitable target," Fu Peng said.
Why has the stock fallen so sharply? At its core, there is too deep a chasm between valuation and fundamentals. In the first half of 2026, Unitree's revenue reached 1.152 billion yuan, up 48.54% year-on-year — respectable on the surface, but growth has slowed sharply from the 332.64% recorded for the full year of 2025. The profit picture looks worse: deducted non-recurring net profit was 244 million yuan, down 19.34% year-on-year. Nomura's target price on listing day was only 370 yuan, nearly 40% below the August 24 closing price of 603.08 yuan. It's not that institutions are bearish on the future of humanoid robots; rather, at the current price, too many future developments have already been priced in ahead of time.
Fu Peng pointed out several realistic issues in Unitree's financial reports. First, the company's end customers are primarily research institutions and universities, with purchases mostly funded by research grants — there is almost no repeat buying, as the cycle ends once funding is secured and procurement is completed. Second, similar robot assembly has become widespread in the Shenzhen region, and as an industry benchmark, Unitree faces considerable competitive pressure. In its response to the second round of inquiry letters, Unitree Technology disclosed that from January to September 2025, its humanoid robot revenue was 595 million yuan, of which 73.6% came from research and education customers, including universities and research institutions. Revenue from commercial consumer scenarios accounted for 17.39% of humanoid robot sales, while industry applications such as corporate tours, smart manufacturing, and intelligent inspection contributed 9.01%. Previously, the share from research and education customers was even higher. In other words, humanoid robot revenue is currently driven by research education, shopping mall tours, and stage performances — selling technology products to research buyers, "selling teaching tools" rather than "selling productivity."
As early as 2025, Zhu Xiaohu, managing partner of GSR Ventures, revealed in an interview that his angel fund had invested in several early embodied intelligence projects over the past few years but had been "exiting over the last few months." He asked the CEOs of these companies the same question: where exactly are your commercialization customers? "I feel like they're all talking about customers they've imagined," Zhu Xiaohu said. "Who would spend over 100,000 yuan to buy a robot to do these tasks?"
Financial commentator Liu Ge told the media that Unitree Technology founder Wang Xingxing himself has publicly stated that embodied intelligence is still very far from its ChatGPT moment. By Wang's own standard — robots must be able to complete 80% of tasks in 80% of unfamiliar scenarios — the current state falls far short. The current high valuation of humanoid robots borrows from Wall Street narrative logic: benchmarking against the global automotive market and imagining humanoid robots entering millions of households, where capturing just a few percent market share would constitute a massive market. "The premise of this valuation model is that humanoid robots can enter households at scale. But the industry reality is that robotic hardware is advancing rapidly while the AI brain remains stagnant. Without a breakthrough in the brain, robots cannot enter homes — they will be confined to specialized scenarios like industry, research, and performance, degenerating into ordinary equipment manufacturing with no essential difference from machine tool companies," Liu Ge said.
In his view, Unitree Technology's reasonable valuation range is 60-100 billion yuan, and general-purpose humanoid robots will not enter households within the next decade. Wang Xingxing himself acknowledges that the "ChatGPT moment" for humanoid robots could take 2-3 years at the earliest, or 5-10 years at the latest, before the industry truly reaches its critical inflection point. But the capital market cannot wait that long. The volatility in Unitree's share price is not a rejection of the humanoid robot direction; it is more like a deep breath after market frenzy. The capital market has already paid a generous premium — now it's up to Unitree to prove itself not merely as an internet-famous robot but as a machine that can genuinely create economic value.
Liu Ge elaborated on why the stock surged so high: multiple factors compounded. First, Unitree Technology is one of the few companies among hundreds of humanoid robot firms that has closed the commercial loop — selling products to research institutions and scenic spots with real shipment volumes, real sales, and actual profitability. This is its biggest fundamental difference from peers. But honestly, that level of profit provides very weak support for the post-listing high share price; the real drivers pushing market value higher were subsequent layers. It is the first humanoid robot stock on the A-share market, carrying inherent scarcity. If the same quality company were listed in Hong Kong or the U.S., its valuation might be cut by five to ten times. But the A-share market needs an embodied intelligence benchmark, and capital captured that signal, joined the fray, and further strengthened market expectations. Finally, after listing, retail investors' enthusiasm for chasing highs pushed the stock up further from institutional pricing levels. The layers compounded to drive it to that height.
On why market value has fallen over 45% from the opening high, Liu Ge pointed to limited application scenarios. The core issue is that the high valuation of humanoid robots borrows from Wall Street narrative logic — imagining humanoid robots entering households at scale, where capturing a few percent of the global automotive market equivalent would be enormous. The premise is mass household adoption. But the industry reality is that robotic hardware is advancing rapidly while the AI brain stagnates. Without brain breakthroughs, robots cannot enter homes and remain confined to industrial, research, and performance scenarios, degenerating into ordinary equipment manufacturing — once the valuation logic collapses, the share price naturally falls.
On whether the "brain breakthrough" is the same concept as the ChatGPT moment Wang Xingxing mentions, Liu Ge confirmed it is the same: the ChatGPT moment refers to a paradigm breakthrough in robot AI brains. Looking at AI history reveals how difficult such breakthroughs are. Early Deep Blue chess victories and later AlphaGo both led people to believe they had found the ultimate path for artificial intelligence, but both proved otherwise; only with ChatGPT did a true large-model paradigm shift occur. Humanoid robots are similar — only with a comparable brain breakthrough can they truly enter households. Why is it so difficult? Current humanoid robot models are fundamentally built on autonomous driving logic. Autonomous driving requires data on the scale of a small town; achieving general-purpose humanoid robots capable of various household tasks requires data on the scale of a solar system. Existing technical routes can progress from a town to a county, but reaching solar-system scale requires order-of-magnitude leaps that cannot be crossed quickly by simply throwing money and people at the problem.
Regarding whether general-purpose humanoid robots can ultimately succeed and when they might enter homes, Liu Ge believes that long-term technical feasibility exists, but his judgment is that general-purpose humanoid robots will not enter households within ten years. The basis for this judgment is the data scale gap and the uncertainty of brain breakthroughs — paradigm shifts are never linear, and you cannot extrapolate the end point from the current progress bar. But this does not mean embodied intelligence is not advancing. Quite the opposite: specialized scenario deployment may happen faster than people think. Just as food delivery robots were already widespread before the so-called ChatGPT moment, outdoor inspection and specific care applications — such as dedicated humanoid robots helping bathe disabled elderly people — may also emerge relatively quickly. China has advantages in the engineering and cost control of specialized robots.
On whether Unitree Technology can be compared to early CATL or BYD and grow into a sector giant, Liu Ge believes the opportunity exists for a giant to emerge in the sector, but the giant may not necessarily be Unitree. First, Unitree's current advantage lies in hardware and dynamic balance, not the brain, and what determines the endgame of humanoid robots is precisely the brain. Second, the brain belongs to the information domain — once technology breaks through, it is easily replicated, and first-mover advantages may not hold. Third, the entire sector is still in its very early stages, far from settled. Looking at internet history, many once-highly-anticipated star companies eventually faded away, while latecomers overtook them. Humanoid robots will likely follow the same path — today's leader is not necessarily the final winner.
On whether the stock still has room to fall after nearly halving, Liu Ge believes it will continue to decline. Its reasonable valuation range is roughly 60-100 billion yuan. This valuation corresponds not to a "general-purpose humanoid robot platform company" but to a "specialized robot and component supplier with stable shipments" — essentially equipment manufacturing valuation logic, not technology growth stock logic. It has real customers and stable shipments, including robot dogs, scenic spots, and university research procurement — these businesses are solid but have low ceilings. Each university globally purchases a few units, limiting market size; scenic spots and performance scenarios cannot support a trillion-yuan company. Having real business and having enormous imagination space are two different things.
On whether Unitree's future leadership in the sector depends on breakthroughs in the "brain," Liu Ge confirmed that the core variable is indeed the brain. But the problem is that the brain belongs to the information domain, where technology is easily replicated and cannot be monopolized behind closed doors by Unitree alone. If a major brain-level breakthrough occurs in the industry, combined with Unitree's already excellent body and dynamic balance hardware capabilities, it would indeed gain wings. To use an analogy, it is like having built a personal computer with excellent hardware, waiting for the Windows operating system to arrive. But Windows may not be developed by you, and once it exists, all hardware manufacturers can install it. So the market remains highly competitive — Unitree has hardware advantages but has not formed an absolute moat. Whether it can transform from "today's leader" to "the final winner" depends on whether it can seize the brain breakthrough the moment it arrives and convert its hardware advantages into ecosystem advantages.
Economist Fu Peng expressed no surprise at the turmoil. Unitree's pre-listing valuation already far exceeded general market consensus, and post-listing volatility was extreme. The robotics industry remains in its very early stages, and despite the completed IPO, it is fundamentally a primary-market venture capital play. Under the fervent expectations of the technology wave, it does not yet belong to long-term value or growth investments. Ordinary investors can speculate on the future with small amounts of capital, but for stable investment returns, it is not a suitable target. The financial reports reveal several realistic problems. First, end customers are primarily research institutions, with purchases funded by research grants — once funds are allocated and procurement completes, the cycle ends. Second, similar robot assembly has become widespread in Shenzhen, and as an industry benchmark, Unitree faces considerable competitive pressure. Additionally, as a benchmark humanoid robot company, the funds raised from listing should be invested in R&D, but its R&D spending is even lower than the R&D investment of Muyuan Foods' pig farming business — this itself signals that the industry is still in its early stages.
On whether Unitree's listing is a typical case of what Fu Peng previously described as the "absorbing star technique" in market conditions, he confirmed it is: on Unitree's listing day, massive market funds were absorbed by it. This is not an isolated case. Looking at the U.S. market over the past six months reveals similar characteristics: the index appears resilient, but the number of stocks effectively supporting the index is rapidly declining. Equal-weight volatility remains elevated while index volatility is suppressed to very low levels — this is the typical structure of capital concentration, where funds increasingly gather into a few core targets while liquidity is drained from peripheral assets.
Among the many heavyweight guests at a recent event, two cross-disciplinary representatives stood out: Xue Qikun, academician of the Chinese Academy of Sciences and winner of the nation's top science and technology award, representing the pinnacle of China's basic research wisdom with his gaze fixed on the technological frontier; and Lin Yuan, a legendary retail investor who grew 8,000 yuan into ten billion in assets, representing the ultimate practice of surviving bull and bear markets with his instincts always attuned to the market's deeper logic.
Section headings have been incorporated into the narrative flow. The article concludes with a brief mention of notable figures attending a recent industry event, which appears in the original source material.
Comments