AsianFin -- If DeepSeek and Unitree Robotics defined the tone of Q1 in China's 2025 venture capital scene, then Q2 is shaping up to revolve around two words: Zhu Xiaohu and humanoid robot bubble.
On March 28, Zhu, the Managing Partner at GSR Ventures, ignited controversy with a blunt remark: "Mass exits from humanoid robot companies."
The comment quickly triggered widespread debate across investment circles, reflecting growing concerns about overheated valuations in the humanoid robotics space.
So, is there really a bubble in the humanoid robot sector? What does a bubble even mean in the context of cutting-edge tech? How big could it get — and what impact will it have on the industry?
As arguments between idealists and realists rage, it may be time to step back and reflect on what "bubbles" truly mean in the worlds of business and technology.
Bubbles and the Business Cycle
The word "bubble" comes from the Latin bulla, meaning something delicate and easily burst — much like its metaphorical usage today.
The term gained economic weight during the Dutch Tulip Mania of the 1600s, when rare tulip bulbs sold for more than luxury Amsterdam homes. Despite warning signs, speculators insisted "this time is different." When prices crashed in 1637, the illusion collapsed — marking history's first recorded financial bubble.
British author Charles Mackay later chronicled this and other speculative frenzies — including the South Sea and Mississippi bubbles of 1720 — in his 1841 classic Extraordinary Popular Delusions and the Madness of Crowds. His work helped cement "bubble" as a catch-all term for irrational market surges.
Over the centuries, bubbles have repeated in strikingly similar patterns: euphoria, overvaluation, collapse. While they often result in financial damage, they have never dulled humanity's appetite for speculation — a dynamic Nobel laureate Robert Shiller and economist Charles Kindleberger attributed to the enduring tug-of-war between greed and fear.
Historically, asset bubbles in goods or financial markets have offered little societal value. But tech bubbles — though also speculative — can lead to meaningful innovation.
Consider the dot-com bubble: Venture investment exploded from $8 billion in 1995 to $104 billion in 2000, fueling infrastructure for cloud computing and e-commerce. Many startups failed, but survivors like Amazon and Google emerged to shape the digital economy.
Or take solar photovoltaics: After the 2000s bust, 80% of global silicon-based solar patents were held by just ten firms, forcing the industry to explore next-gen thin-film technologies.
The autonomous driving bubble saw players like Waymo and Cruise pour over $100 billion into R&D, accelerating advances in lidar and AI perception.
Even the 2018 blockchain crash left a legacy: a 300% spike in crypto developers, 35% of whom came from traditional finance. That influx of talent laid the groundwork for innovations like Web3 and DAOs.
In each case, the bubble's burst left pain — but also progress. As science fiction writer Liu Cixin once said, "Bubbles are the entropy cost of civilization's upgrade."
Speculation vs. Innovation
Of course, not all bubbles are created equal. Today's financial elite often blur the lines between arbitrage-driven hype and genuine technological innovation. Grand narratives of progress are sometimes used to mask short-term trading goals. But even so, this dynamic isn't necessarily bad: idealists envision the future, and speculators help accelerate its arrival.
This duality is embodied in every venture capitalist. They must champion the next breakthrough while managing the risks of hype-driven market dynamics. Their goals? To create waves and time the peak.
How do you spot a bubble? How do you gauge its size? How do you ride the wave — and exit before the crash?
There's no definitive playbook. Each investor relies on their own experience, instincts, and beliefs. Yet as humanoid robotics enters the spotlight, these questions are becoming central to the venture capital playbook.
For entrepreneurs and investors alike, understanding the anatomy of a bubble may be just as crucial as building the next big thing.
Warning Mechanism
Venture capitalists have compiled a multi-dimensional warning model from numerous public sources, summarizing the insights of top global venture capital firms and consulting companies regarding industry bubbles. This model includes eight key indicators to define the emergence of industry bubbles and assess bubble pressure.
1. Total Asset Indicator: The short-term growth in the number of enterprises within the same niche sector.
In 1995, the IPO of Netscape Navigator triggered a frenzy of capital chasing the internet, leading to the establishment of over 10,000 new .com companies in the U.S. between 1998 and 2000.
During the group-buying bubble from 2010 to 2013, over 20,000 companies were registered globally, with approximately 12,000 in China, primarily focused on local lifestyle services.
Between 2014 and 2016, during the shared economy bubble, over 6,000 companies were registered globally, including 1,800 in China.
Sequoia Capital highlighted in its Risk Factor report (2021) that when the annual growth rate of companies in a particular sector exceeds 200%, one should be cautious of irrational exuberance. Similarly, McKinsey emphasized in its study The Rise and Fall of Industries (2022) that when the growth rate of companies in an industry is more than five times the GDP growth rate, the risk of a bubble significantly increases.
2. Funding Heat Indicators:
Growth rate of total funding in the sector.
The growth of total funding in a sector is often the direct cause of a simultaneous surge in asset volume. For instance, during the shared economy bubble in 2015, the global funding total for this field exceeded $20 billion, marking a year-on-year growth of 120%. The following year, the funding total for China's shared bike market exceeded $9 billion, with a quarter-on-quarter growth rate of 200%. In SoftBank's 2019 earnings call, Masayoshi Son mentioned that when quarterly funding in a sector grows by more than 300% quarter-on-quarter, it is necessary to reassess asset quality.
Proportion of angel round funding.
A high proportion of angel round funding indicates a large influx of startups into the sector. The inherently low survival rate of startups is itself a major factor contributing to the bursting of bubbles. In 2022, the proportion of angel rounds in the metaverse sector reached 41%, with valuations evaporating by 67% within six months. Benchmark Capital founder Bill Gurley noted in his 2022 Web3.0 investment memo that when the proportion of angel rounds exceeds 30%, it is a sign that the industry has entered an irrational phase.
Compression of funding rounds.
Another metric indicating the rapid inflow of capital. During the O2O bubble in 2015, the average interval between seed rounds and Series B rounds shortened from two years to six months. Similarly, during the global ICO bubble in 2017, the average interval between seed rounds and PE rounds shrank from six months to one month. Goldman Sachs' 2015 report The Rise of the Mega-Rounds noted: "During bubble cycles, the interval between funding rounds significantly shortens. For early-stage projects (e.g., seed rounds to Series A rounds), the average time compresses from the traditional 18–24 months to 6–12 months, reflecting overheated capital and intensified competition."
FA commission premiums.
The fee standards for financial advisors (FA), as the data indicator closest to transactions, can also be used to gauge the level of capital enthusiasm for a particular sector, and it provides timely insights. According to PitchBook statistics, FA fee rates in the VR sector increased by 20%-30% year-on-year in 2021.
The FA Industry White Paper (2020) pointed out that when FA fee rates exceed 5% and are accompanied by exclusivity clauses, the industry is likely in an overheated state.
3. Irrational Pricing Indicator: Price-to-Sales (PS) Ratio:
This indicator is often criticized by analysts in the secondary market for highlighting asset pricing bubbles in the primary market. For example, in the global SaaS bubble of 2021, the PS ratios of several unprofitable SaaS companies reached as high as 30x, while Adobe's PS ratio was only 10x. Morgan Stanley also issued a warning in its Tech Bubble 2.0 report, stating that caution is needed when SaaS companies' PS ratios exceed 30x.
Benchmark Capital founder Bill Gurley also pointed out that when a company's PS ratio exceeds 40x, there is a risk of "unit economics failure."
4. Exit Channel Indicator:
According to data from The Wall Street Journal, the first-year post-listing break rate for SPAC-listed companies on the U.S. stock market reached as high as 63% in 2021, earning the label of the U.S. SPAC bubble. Alongside the post-listing declines, a significant drop in merger and acquisition activity in the sector also signals the arrival of a bubble.
Goldman Sachs statistics show that when the first-year post-listing break rate exceeds 60%, the market enters the "bubble deflation stage."
5. Capital Attribute Indicator: Influx of Cross-Sector Investors
This is characterized by the large-scale entry of traditional industry giants. Notable examples in recent years include the pre-burst phase of the electric vehicle bubble: Baidu established Jidu Auto, Evergrande Group acquired Faraday Future (FF) and founded Evergrande Auto, Baoneng Group acquired Qoros Auto, and China Fortune Land Development acquired Hozon Auto to launch Nezha Auto, among others.
Bain & Company, in its Corporate Venturing 2022: Scaling Innovation Through Collaboration report (2022), conducted a detailed analysis of the cyclical characteristics of corporate venture capital and mentioned: "When cross-sector capital (especially traditional industry capital) accounts for more than 30% of a new emerging field, the average valuation deviation in that field increases by 40%, and the probability of industry reshuffling within the next three years exceeds 60%."
6. Talent Magnet Indicator: Irrational Surge in Salary Levels
During the ICO frenzy of 2017, blockchain startups widely offered high salaries to attract talent, with annual salaries for ordinary programmers reaching $300,000–$500,000 (far exceeding the average levels in Silicon Valley). According to LinkedIn's Salary Insights report, global salaries for metaverse-related positions increased by 120% year-on-year in 2022, yet 90% of the projects remained at the conceptual stage.
The influx of short-term speculative capital drove up labor costs, causing salary growth to far outpace industry revenue or profit growth, significantly deviating from fundamentals—an indicator of an impending bubble burst.
7. Narrative Heat Indicator: Media Hype Index
Before Bitcoin's crash in 2018, global mainstream media (such as The New York Times and The Wall Street Journal) featured daily front-page reports, and search volumes for "Bitcoin" on social media surged by 300%. Similarly, on the eve of the "metaverse" bubble burst in 2021, Google Trends showed a 700% month-on-month increase in searches for the keyword "metaverse." According to statistics from the Zero2IPO Research Center, the peak of Baidu Index searches for the O2O sector in 2015 coincided with its peak financing period with a 92% overlap. Media overhype often acts as an accelerator for bubbles, while a shift to negative narratives exacerbates market crashes.
CB Insights' Narrative Analytics tool uses metrics such as media mentions and social media discussion heat to predict bubble risks.
8. Regulatory Behavior Indicator: Policy Arbitrage and Frequency of Regulatory Guidance
This indicator is relatively unique. For instance, in certain "quasi-financial industries," before explicit bans were imposed, over 70% of projects relied on regulatory gray areas for their business models. Similarly, in the "gardener track," before facing regulatory crackdowns, the average number of monthly meetings with regulatory authorities exceeded three.
According to Ernst & Young's research report Global Regulatory Outlook 2023, compliance costs in heavily regulated sectors have been growing annually by 150%–200%.
What Venture Capitalists Says
Simply inserting industry data does not directly lead to conclusions about bubbles. Just as we cannot predict the trajectory of the 20th-century internet bubble using the price fluctuations of 17th-century tulip bulbs, we also cannot apply the financing cycle compression patterns of the Web 2.0 era to AI large models or humanoid robotics. Changes in the macroeconomic environment, adjustments in policy frameworks, and disruptive technological revolutions continuously reshape the soil for bubble formation and the tipping points for their collapse.
The "300% Quarterly Financing Growth Threshold" outlined in the Goldman Sachs report and the "30% Angel Round Warning Line" set by Benchmark Capital may serve as precise warning signals today, but tomorrow they could become outdated maps reconstructed by new variables.
It is precisely this dynamic nature that gives the warning mechanism its true vitality. When we attempt to use data to anchor bubbles, we must also acknowledge the inherent uncertainty behind them. Thresholds are fluid, and true wisdom lies in sensing change.
Historical experience shows that truly disruptive innovations often emerge on the shores after the tide of bubbles recedes. When the wave of capital subsides, only those companies with strong technological moats and clear commercialization pathways can endure through cycles, transforming today's bubbles into tomorrow's cornerstones. The bubble itself is not the problem; misjudging the timing of its burst is what proves fatal.
