Hangzhou (杭州). 10:42 a.m.
Wang (pseudonym, age 31) sits on an electric scooter, checking the order notification on his Meituan (美团) app. The destination is an apartment complex 1.8 kilometers away, fourteenth floor. Estimated delivery time: eighteen minutes. The app's algorithm has already calculated the optimal route. All he has to do is follow it.
His degree is a master's in computer science.
His 2022 thesis was on load-balancing algorithms for distributed systems. That year, he applied for a new-graduate position at Alibaba (阿里巴巴). He passed the first technical interview. The algorithm coding test went smoothly. He was cut in the second round. The interviewer never told him exactly where he fell short. He applied again at Tencent (腾讯). Screened out at the resume stage. Applied to Baidu (百度). No response. In 2023, and again in 2024, he repeated the same process. Even internship openings were shrinking. The job market had been frozen for two years.
That summer, he downloaded a delivery app.
It was not just about the money. Riding a scooter through the city felt less desperate than sitting at home in front of a monitor, revising his resume for the hundredth time. At least when your body is moving, you retain the sensation of doing something.
His monthly income is 6,800 yuan (元/CNY). About $950. His parents pay 12,000 yuan a month on the mortgage for their Hangzhou apartment. Even if Wang works every single day without rest, his earnings barely cover half of his parents' monthly loan payment.
That apartment requires its own discussion.
His parents bought it in 2020 for 5 million yuan. It was near the Hangzhou tech cluster, two subway stops from Alibaba's headquarters. The Hangzhou tech boom was at its peak. The calculation was straightforward: "Once our son lands a job at a tech company, this neighborhood will be perfect." They put every asset the family owned into the mortgage as collateral.
By 2025, the apartment's market value is 3.5 million yuan. The outstanding loan balance is 3.8 million yuan. Selling it would not cover the debt. Sell and you lose. Hold and you bleed interest. This is what the Chinese call fu zichan (负资产) — negative equity. The home is no longer an asset. It is a liability.
Wang does not use the word tang ping (躺平). He just says he is living. He listens to Python tutorials through his earbuds while making deliveries. He believes he will return to a coding job someday. Or perhaps what he believes is that he wants to believe it. The app sends the next order notification. 1.2 kilometers. Twelve minutes. He rides off.
Wang's story is not individual misfortune.
The age-35 crisis (sanshiwu sui weiji, 35岁危机), graduate unemployment, the real estate collapse, the demographic cliff. These appear to be separate events, but they are a structural chain — each one reinforcing the others. There are millions of people like Wang across China. Some 240 million workers participate in the gig economy (National Bureau of Statistics, 2025). Delivery riders alone number over 10 million. Among the riders, college graduates account for a significant share. In 2023, the claim that "tens of thousands of Meituan riders hold master's degrees" circulated on social media; Meituan officially denied it (美团, 2023). Precise education-level statistics have not been released. But given the overall scale of the rider workforce and the severity of the graduate employment crisis, multiple field reports confirm that highly educated riders are far from rare.
Their collective choices are reshaping the macroeconomy.
In this chapter, we encounter China's displaced across four layers. The first layer is the cliff at age 35 — a system that extracts maximum labor during youth through 996 culture, then discards the spent workforce at around thirty-five. The second layer is tang ping and bai lan (摆烂) — the refusal to participate, understood as rational economic behavior. The third layer is the real estate crisis of negative equity — the landscape in which the very identity of the middle class is disintegrating. The fourth layer is the demographic cliff — a future already locked in, and the paradox hidden within it.
We trace how these four layers interlock.
Section A: The Cliff at 35 — Institutionalized Age Discrimination
The Numbers Speak First
In 2021, Baidu (百度) had 45,498 employees. By the end of 2024, the number was 35,900. A 21.1% reduction in three years.
Alibaba's trajectory is more dramatic. In March 2022, 254,941 full-time employees. By March 2025, 124,320. A 51.2% cut. The world's largest e-commerce company shed half its workforce in three years. It cut headcount on this scale while its services kept running. AI and automation filled the gaps.
An unwritten practice has long operated inside China's tech industry. It appears in no official document, yet everyone knows it. The age-35 crisis (35岁危机) works like this: when restructuring is needed, employees over 35 are the first targets. When an HR meeting gets scheduled, those who receive the invitation already know the outcome. Some pack their things in advance.
Why 35?
Three structures converge on the same point.
The first is the salary curve. In Chinese tech companies, the pay gradient by seniority is steep. By age 35, an employee's total compensation runs at 2.5 times or more that of a junior hire. But management positions are scarce. In a pyramid, not everyone can move up. When seniority costs accumulate, companies run the numbers.
The second is the paradox of 996 culture. Arrive at 9 a.m., leave at 9 p.m., six days a week — this regime is a machine for extracting peak output from the young. In the late twenties to early thirties, when stamina and focus are at their height, companies push employees to sustain this intensity. In return, the companies get rapid growth. But when the same person hits 35, the body sends different signals. Sleep shortens. Recovery slows. The company swaps in fresh labor. The structure is simple: burn through the young under 996, and when they burn out, slot in the next cohort.
The third is the acceleration of AI. AI coding tools have begun replacing the work of junior engineers and mid-level data analysts at speed. The result: the pathway for mid-career professionals around 35 to advance into senior roles has narrowed. AI presses upward from below; a ceiling blocks movement from above. Thirty-five is the point where both pressures converge with the greatest force.
And there is no institutional safeguard.
The Critical Difference from the United States
The United States has the Age Discrimination in Employment Act (ADEA), enacted in 1967. Firing or refusing to hire an employee aged 40 or older on the basis of age is a federal offense. The law does not catch every case. In 2025, direct AI-related layoffs in the U.S. reached 55,000, and tech-sector layoffs in January and February 2026 alone totaled 32,000. America has its displaced too.
But America's displaced operate within a legal perimeter. They can file lawsuits. They can gather evidence. They can negotiate settlements. The difference is not that every negotiation succeeds — it is that the negotiating table exists at all.
In China, it effectively does not.
Labor law formally prohibits age discrimination, but the age-35 crisis is a practice that officially does not exist. When a company lists "deterioration of business conditions" as the restructuring rationale, it faces no legal exposure. For employees, proving age discrimination is functionally impossible. Labor dispute mediation bodies treat the issue with passivity. This is how a practice hardens into a norm.
Signs of change are not entirely absent. In the 2025 national civil service examination, the age ceiling for certain positions was relaxed from 35 to 38 or 40 (National Civil Service Administration). The government indirectly acknowledged the problem with the 35-year threshold. But there is no binding authority over private enterprises. One line changing in a civil service exam does not change the resume filter in a tech company's HR department.
According to Pei Xiaomei (裴晓梅), a sociologist at Tsinghua University, private companies have begun using the age of 35 itself as a pretext for dismissal and hiring restrictions. A 2025 empirical study by Wu Fan (吴凡) at the Chinese University of Hong Kong proves this in numbers: the tipping point at which unemployment risk spikes shifted from age 54.3 in 2018 to age 40.3 in 2022 — a fourteen-year acceleration in just four years. Age discrimination is transforming from social convention into statistical fact, and the speed of that transformation is striking.
How this practice operates is laid bare in the words of those on the receiving end. Jiang Weicheng (江玮城, age 35) was an HR supervisor at an internet e-commerce company. Working in recruitment, he learned the hidden criteria. "There is an age standard concealed at the resume screening stage (在简历筛选环节，隐藏着一个年龄标准)." Department heads were filtering out applicants between 30 and 40. When Jiang himself left the company and entered the job market in January 2024, he experienced the reality firsthand. "Three or four out of every ten job postings explicitly state 'under 35' (10个岗位里至少有三四个直接写了35岁以下)."
Zhou Hao (周浩, age 36) was a project manager in financial IT. At a job interview, the recruiter told him directly: "We don't take anyone over 35. No matter how talented (35岁以上的不要，不管他多优秀)."
Zhang Yang (张洋), a 25-year-old internet headhunter, summarized the industry's unwritten rule: "Thirty-five is the common denominator across every company. If you haven't reached 1.8 million yuan in annual salary by 35, you're getting cut (35岁是各个公司的一个公约数...如果你35岁没到180万，你是要被开掉的)." If the average age at a major internet company is 30 to 32, where did the people over 35 go? (The Paper, 2022).
Caixin (财新) traced the personal consequences of this structure in an August 2025 investigative report. Xu Yang (徐阳, age 46) was laid off from the China branch of a U.S.-based multinational software company as U.S.-China trade tensions escalated. The company cut 70% of its China-based workforce. Xu was 46 — well past the wall of the age-35 crisis. Returning to a comparable position was practically impossible. The same Caixin report found that the share of flexible-job postings rose from 8.4% in 2019 to 15.2% in 2024 — nearly doubling. Permanent positions are shrinking; non-standard employment is expanding. A structural transition is underway.
The same AI shock is hitting both countries, but the institutional environments in which it operates are different. America's displaced are pushed within the perimeter of legal protection. China's displaced are pushed in a state of structural defenselessness. This is what creates the difference in speed and intensity.
Kuaishou's "Limestone" Project
Second half of 2024. Inside Kuaishou (快手), a restructuring initiative codenamed "Limestone" (石灰石) was set in motion.
It was a Tuesday afternoon. At Kuaishou's Beijing headquarters, several employees in their mid-thirties on a development team received sudden HR meeting notifications. The notices arrived via company messenger. A conference room number and a time. Nothing else. Those who received the message already knew what conversation awaited them on the other side of the door.
The message delivered in the conference room was clear. Your contract is terminated effective today. You have two hours to pack your belongings. A compensation package was placed on the table: N+1, the standard Chinese statutory severance formula. N equals years of service; +1 is one month's salary. The document included a clause: by signing, you agree not to pursue further legal action.
Packing took less than thirty minutes. Laptops had to be returned. Personal belongings filled a single cardboard box. Walking out of the building, one employee pulled out his smartphone. He opened the Meituan app for the first time. He checked the registration process for delivery riders. That was the same night.
This scene has repeated thousands of times. Only the company name changes; the structure stays the same. Alibaba, Baidu, ByteDance (字节跳动), JD.com (京东). The tech giants moved in the same direction simultaneously. Hiring froze. Existing headcount shrank. AI tools filled the gaps.
A Tectonic Shift in Graduate Employment
If the age-35 crisis operates from above, the new-graduate market exerts simultaneous pressure from below.
In the first half of 2025, job postings for college graduates fell 22% year on year. The pace of hiring cuts far outstripped the rate at which graduates were entering the market. Demand contracted while supply expanded.
Graduates read the signal fast. In a 2024 employment preference survey (Zhaopin data), 47.7% of college graduates selected state-owned enterprises as their top choice. Private-sector companies drew just 12.5%. Five years earlier, private-sector preference exceeded 40%. The reversal is complete.
These figures say two things at once.
First, young people have internalized the instability of the private sector. They watched Alibaba shed half its workforce in three years. They read about Kuaishou's "Limestone" on social media. They confirmed, through data, that the era in which landing a job at a major private company was a badge of success has passed. It is not that state-owned enterprises are especially attractive. It is that private companies are more dangerous.
Second, talent is exiting the innovation ecosystem. The rush toward state-owned enterprises is a flight to safety and simultaneously a retreat from entrepreneurial ambition. The private tech sector, the engine of Chinese innovation, is losing access to top talent. The consequences for China's innovation ecosystem are still being written.
Two pressures act simultaneously. From above, the age-35 crisis drains experienced workers. From below, new talent is absorbed into state-owned enterprises. The private tech sector is losing human capital from both directions at once.
Section B: Tang Ping and Bai Lan — Lying Flat and Letting It Rot
Living on 2,000 Yuan
One day in 2023, a post appeared on Baidu Tieba (贴吧) and was deleted. But screenshots were already circulating through WeChat (微信) group chats.
It read:
"I live on 2,000 yuan a month. Rent, 800 yuan. Food, 600 yuan. Phone bill, 200 yuan. Remaining, 400 yuan. I'm not studying for the kaoyan (考研, graduate school entrance exam). I'm not sitting for the civil service exam. I'm not doing 996. I'm sorry to my parents, but I've realized this game cannot be won. I've decided not to spend my energy on a game I cannot win. That's all."
The author was believed to be a 27-year-old man. He had graduated from a teachers' college and failed the teacher certification exam twice. He had no motivation to try again. He covered his living expenses with a convenience store job in a cramped studio on the outskirts of the city. Two thousand yuan. Not enough, but the minimum.
The post was deleted. But the sensation he expressed already belonged to millions.
Tang Ping: An Economic Interpretation
Tang ping (躺平) is an expression that spread across the Chinese internet in 2021. Literally translated, it means "to lie flat." It is a declaration of refusal — a rejection of high-intensity competition and 996 work culture, an embrace of minimized consumption and bare subsistence.
Bai lan (摆烂) takes it one step further. The literal meaning is closer to "let it rot" or "let it fall apart." If tang ping says "I will live at the minimum," bai lan says "I will no longer try to improve the situation." It is the active acceptance of decline.
To read this as generational laziness is to get it wrong.
Through an economic lens, the picture looks different. In a game where expected costs exceed expected returns, ceasing to participate is rational behavior. You cannot buy a home even working 996. You cannot save enough for marriage. You cannot expect a secure retirement. When the question "Why should I work 996?" has no convincing answer, the rational actor reduces participation in the system.
Tang ping and bai lan are the collective conclusion.
The Closed Doors, in Numbers
As of August 2025, the youth unemployment rate (ages 16–24, excluding students) stands at 18.8%, the highest since the revised methodology was introduced.
Behind this figure lies a backstory. When the National Bureau of Statistics (国家统计局) recorded a youth unemployment rate of 21.3% in 2023 (a record high), it suspended publication of the statistic altogether. It later resumed reporting under a new methodology that excluded enrolled students. Even under that revised methodology, the rate is 18.8%. Newsweek estimated the number of unemployed urban Gen Z workers at 20 million.
The graduating class of 2026 is projected at 12.7 million, the largest in history. The year these graduates enter the job market coincides precisely with the escalation of the U.S.-China trade war and the full onset of the AI shock.
When one door is shut, you knock on another. Two alternative pathways have emerged.
Applicants for the kaoyan (考研, graduate school entrance exam) numbered 3.88 million in 2025, more than double the figure from a decade earlier. Not all of them are driven by scholarly ambition. The job market is blocked, so they stack another degree. It is the graduate-school equivalent of cramming for a retake. But graduate school seats are finite too. As competition ratios climb, the number of rejects rises in tandem.
Applicants for the national civil service exam (国考) reached 3.41 million in 2024. The average competition ratio was 77 to 1. Seventy-seven people apply; one gets the job. For popular positions, the ratio exceeds several hundred to one. This is the last exit to safety.
Competition explodes while openings do not expand. When the structure of the game looks like this, some players refuse to play at all. Tang ping is the language of that refusal.
Tang Ping Is Not Unique to China
Reading this as a peculiarity of Chinese youth psychology is a misreading.
South Korea's "N-po generation" (N-po sedae) refers to a generation that has given up on dating, marriage, childbirth, and homeownership. Japan's satori generation has surrendered desire itself. When the Great Resignation swept the United States in 2021, millions of Americans reconsidered the employment relationship altogether.
The structure is the same. When expected returns decline amid high-intensity competition, the path that rational actors collectively choose is similar regardless of nationality.
China's distinctiveness lies in two dimensions. The first is scale. When hundreds of millions out of a population of 1.4 billion make the same choice, the macroeconomic structure shifts. The second is the narrowness of institutional alternatives. In the United States, leaving a large corporation opens the door to a startup or freelance work. In China, leaving the private sector means lining up for a state-owned enterprise exam or descending into the gig economy.
When tang ping youth cut spending, domestic demand contracts. When they postpone marriage, the birth rate falls. When they abandon entrepreneurial ambition, the startup ecosystem weakens. Rational individual choices, aggregated, become macroeconomic costs.
The Government's Response — and Its Paradox
The Chinese government perceives tang ping as a political threat.
The state narrative is built on fendou (奋斗, struggle and striving) and the "great rejuvenation of the Chinese nation" (中华民族伟大复兴). The collective disengagement of the young directly negates that narrative. In 2021, Xi Jinping (习近平) declared that youth must "pursue a life of struggle," an indirect rebuke of tang ping culture. In 2025, the Cyberspace Administration of China (CAC, 国家互联网信息办公室) issued directives to censor tang ping-related expressions on social media.
Censorship amplifies the very thing it targets. Deleted posts spread further as screenshots. Erasing a word from a platform does not change the underlying reality when the word has already entered everyday speech. The content of that Tieba post — long deleted — still circulates in WeChat group chats today.
Section C: The Generational Shock of the Real Estate Collapse
With What Money?
Hangzhou. A weekend in the fall of 2025.
Mr. Liu (pseudonym) and his wife (early thirties, pseudonym) bought a 5-million-yuan apartment when they married in 2020. It was near the Hangzhou tech cluster. Both families pooled their assets and took on debt. At the time, marriage and homeownership came as a package deal. Social pressure dictated that without a home, marriage was difficult.
The husband's monthly salary is 15,000 yuan. Of that, 12,000 yuan goes straight to the mortgage. Two people live on 3,000 yuan. They almost never eat out. The last time the couple traveled together was two years ago.
By 2025, the apartment's market value is 3.5 million yuan. The outstanding loan balance is 3.8 million yuan. Sell and you still owe. Hold and the interest compounds. And there is the fear that the market value could fall further.
A journalist asked whether they planned to have children.
The wife smiled. "With what money?"
It was not anger. It was not resignation. It was arithmetic. Given the numbers of their current situation, the answer was self-evident.
A City of Collapsed Assets
Shenzhen (深圳) is the front line of China's real estate collapse.
Shenzhen's Futian (福田) and Nanshan (南山) districts are premium business zones where Tencent, Huawei, and major financial institutions are concentrated. At the 2021 peak, luxury apartments in these areas exceeded 100,000 yuan per square meter. As of 2025, asking prices for the same properties have fallen 30–40% from the peak. Some distressed sales show even steeper declines. An investment of 10 million yuan is now worth 6 to 7 million.
Across the top 100 cities, secondhand housing prices fell 7.60% year on year as of October 2025. The number alone may not look dramatic. But when you consider that 70–80% of Chinese household wealth is concentrated in real estate, it takes on an entirely different meaning.
Why this extreme concentration? There were no alternatives. China's stock market is volatile. Bank deposit rates have often fallen below the rate of inflation. Overseas asset transfers are tightly restricted by foreign exchange controls. Real estate was effectively the only credible investment vehicle. For twenty years, Chinese property prices moved in one direction: up. On that belief, tens of millions of households built their financial plans.
That belief is collapsing at the same speed as prices.
Real estate investment in the first three quarters of 2025 fell 13.9% year on year. Goldman Sachs estimated that the real estate downturn dragged China's real GDP growth rate down by 2 percentage points annually in 2024–2025. Non-performing loan (NPL) rates on mortgages at listed banks rose by more than 20 basis points in the first half of 2025.
What happens inside a negative-equity household is straightforward. The bulk of income is locked into loan repayment; disposable income vanishes. National Bureau of Statistics (NBS) data for 2025 reveals the structure: household income rose 5.1% year on year, but consumer spending increased by only 4.6%. Households are not spending as fast as they earn. The IMF identified the negative wealth effect from real estate as one structural cause of China's persistently high household savings rate (IMF Working Paper, December 2025). Consumption dies. When consumption contracts, domestic demand weakens. When domestic demand weakens, the economy deteriorates. When the economy deteriorates, the job market freezes further. A frozen job market threatens the next generation's ability to service their loans. The cycle is self-reinforcing.
Different Generations, Different Shocks
Each generation absorbs the shock in a different shape.
Those born in the 1980s are now in their early forties. They were the most active property buyers during 2005–2015, the generation that entered at mid-range prices. Today they face the age-35 crisis and real estate losses simultaneously. They have reached the age at which tech companies target them for restructuring, and their homes have slipped into negative equity. Two shocks overlap.
Those born in the 1990s are now in their thirties. They are the hardest hit. Between 2016 and 2021, they bought at peak prices, often tying the purchase to marriage. Entire families pooled debt. The largest financial decision of their lives coincided with the worst possible timing. The Liu couple in Hangzhou belong to this generation.
Those born in the 2000s are now in their twenties. Falling prices should, in theory, work in their favor; homes are cheaper than they were at the 2021 peak. But they cannot find jobs. Without employment, they cannot qualify for a mortgage. They are the generation that cannot afford to buy the homes that have become cheaper.
Three generations are pinned inside the same structure in three different ways.
The Collapse of Belief Runs Deeper
The true shock of the real estate collapse lies behind the asset-loss figures.
Through the early 2020s, a belief held urban middle-class China together: work hard, buy property, and your wealth will grow. This narrative had run unbroken for a generation since the Reform and Opening Up of the 1980s. That belief was embedded in the physical form of real estate. An apartment was the most reliable thing a parent could pass on to a child.
That belief has collapsed.
They worked hard. They bought property. Their wealth shrank. The loan remains. This experience hit tens of millions of households simultaneously. Consumption fell. Births stopped. Trust in government eroded. The real estate collapse is not an asset crisis. It is the collapse of a social contract.
And the debris of that social contract collides with tang ping and bai lan culture. The recognition that "no matter how hard I work, I can never catch up to my parents' generation of property wealth" leads to the abandonment of effort, the abandonment of consumption, the abandonment of marriage, the abandonment of childbirth. Each act of abandonment connects to the next, reinforcing the structure.
Section D: The Demographic Cliff — A Future Already Locked In
The Journalists Did Not Ask Questions
January 17, 2026. A press conference at the National Bureau of Statistics (国家统计局) in Beijing.
A bureau director stepped to the podium. This was the 2025 population data release. His voice was measured. He read from a prepared script.
Births in 2025: 7.92 million.
A 17.2% decline from the previous year. The lowest figure since the founding of the People's Republic of China in 1949.
The press gallery was silent. No hands went up. No questions were asked. Everyone in the room understood what the number meant.
Walking out of the briefing room, one journalist said to a colleague: "Those 7.92 million have already determined the labor market of 2045. And the consumer market of 2055."
Demography is the most predictable variable in social science. A child not born today will be absent from the labor market in twenty years. Absent from the consumer market in thirty. Not a taxpayer in forty. There is effectively no policy instrument that can alter this locked-in future.
A Succession of Historic Lows
The 7.92 million must be placed in context.
China's annual births stood at 17.86 million in 2016. In fewer than ten years, the number has fallen by more than half. Four consecutive years of decline: 10.62 million in 2021, 9.56 million in 2022, 9.02 million in 2023, 9.54 million in 2024 (a brief uptick driven by the Year of the Dragon effect), 7.92 million in 2025. The drop from 9.54 million in 2024 to 7.92 million in 2025 was particularly steep: 1.62 million fewer births in a single year.
The crude birth rate is 5.6 per thousand. The total fertility rate (TFR) stands at 1.0 births per woman, less than half the replacement level of 2.1. Beijing and Shanghai register 0.6 to 0.7, among the lowest in the world.
Total population has declined for four consecutive years to 1.405 billion. The net decrease in 2025 alone was 3.4 million, the largest drop since the Great Famine of 1959–1961.
The share of the population aged 60 and above reached 23% in 2025. By 2035, it is projected to exceed 30%. A society in which some 420 million people are elderly. The number of dependents each working-age person must support is rising fast.
The Chinese government has deployed every policy instrument available. In 2015, it switched to a two-child policy. In 2021, it moved again to a three-child policy. Subsidies were offered. Tax deductions were introduced. Housing incentives were rolled out. The birth rate continued to fall.
Why have the policies failed? Because declining fertility is not simply a matter of insufficient economic incentives. The values of highly educated women regarding marriage and childbirth have shifted. The cost of urban life is too high. Education and housing expenses for raising a child are unaffordable. The causes are entangled: real estate collapse, job market instability, and tang ping culture reinforcing one another. The experiences of Japan and South Korea had already foreshadowed this outcome. A shift in values has occurred — one that economic incentives alone cannot reverse.
The Trap of "Getting Old Before Getting Rich"
Japan confronted its aging society when per capita GDP had already surpassed $30,000. It entered demographic transition with a robust social safety net already in place.
China is different. It is entering hyper-aging with a per capita GDP of $13,000. The Chinese call this wei fu xian lao (未富先老) — "getting old before getting rich."
The pension system, medical infrastructure, and social safety net have not been fully built before the burden of elder care explodes. Already, thirty-somethings whose income is 80% consumed by mortgage payments on negative-equity homes now face the added cost of supporting aging parents. The state asks young people who are locked out of the job market to financially support their parents. This is the final layer of the tang ping structure.
Cai Fang (蔡昉), an academician at the Chinese Academy of Social Sciences (中国社会科学院), laid out this structural contradiction in a May 2025 public lecture: "The impact of the current technological revolution on employment is unprecedented in speed, breadth, and depth. The temporal asymmetry between the speed at which technology displaces jobs and the speed at which new jobs are created — along with the structural contradictions — cannot be resolved by simply 'remaining optimistic and criticizing Luddism.'" — Cai Fang (蔡昉), Chief Expert of the National High-End Think Tank (国家高端智库), Chinese Academy of Social Sciences, May 2025 (reported by Tencent News).
The Paradox: Population Decline Accelerates AI
Here, an unexpected structure emerges.
When the population shrinks, the labor force contracts. Wage pressure builds, especially in manufacturing and logistics. When wages rise, companies gain an incentive to invest in automation. When automation investment increases, AI adoption accelerates.
China leads the world in industrial robot installations as of 2023 (270,000 units per year, accounting for 70% of global deployment). Robot density — the number of robots per 10,000 manufacturing workers — stands at 470, third-highest in the world. These figures represent both a commitment to the "world's factory" identity and a response to the demographic cliff.
The paradox completes itself here.
College graduates who have descended into delivery work coexist with robots arriving in factories to replace them. Technology eliminates jobs on one front. On the other front, the resulting labor shortage justifies even faster automation.
Wang listens to Python tutorials through his earbuds while making deliveries. That Python could someday be used to write the operating algorithm for delivery drones. Those drones could eliminate Wang's job. The same technology operates in two directions simultaneously. It pushes some people out, moves into the space they vacated, and uses the reason for moving in to justify moving in further.
This is China's Engels' Pause. Just as productivity rose in Britain between 1780 and 1840 while real wages stagnated for decades, in China too, the gains from productivity improvement are failing to reach labor. How long this period will last remains unknown.
Volume 1 Connection: Roman Smallholders and China's Gig Workers
In Volume 1, we analyzed the collapse of the Roman smallholder.
In the second century BCE, two consequences of Rome's wars of conquest squeezed the small farmer. Slave labor captured in war destroyed the cost structure of agriculture. Capital extracted from conquered territories expanded the latifundia, the great estates of the landed elite. Smallholders could not compete with this cost structure. They sold their land and migrated to the cities. They became the proletarii — free citizens without property.
As the smallholders collapsed, so did the foundation of Rome's citizen army. At the time, only citizens holding a minimum level of property were eligible to serve. The propertyless were not soldiers. As the ranks of the proletarii swelled, the citizen army shrank.
The military reforms of Gaius Marius in 107 BCE filled the gap. He abolished the property requirement and enlisted the landless. It was the beginning of professionalization — and mercenarization. The loyalty of the legions began shifting from "Rome" to "the individual commander." The civil wars of Caesar and Pompey grew from that structure. The fruits of conquest had, paradoxically, driven out the conquerors.
China's 2020s are replicating this structure.
In the 2000s, China's manufacturing boom drew rural youth into factory work. They powered the world's factory. Their labor created China's economic growth. It was the fruit of conquest.
In the 2010s, automation and rising labor costs pushed those workers out. Factories brought in robots. Low-wage manufacturing jobs disappeared. Workers descended to the next level down. The gig economy. Delivery riders. Platform day labor.
In the 2020s, AI and platform automation began pushing out the next layer: college graduates. A computer science master's student becomes a delivery rider. The gig economy encompasses some 200 million participants. Delivery riders alone exceed 10 million. Precise educational statistics remain undisclosed, but field reports consistently indicate that highly educated riders are growing rapidly in number.
As Roman smallholders became the proletarii, Chinese graduates become gig workers.
The migration of loyalty follows a parallel track. Just as Roman smallholders who became proletarii shifted their identity from "citizen of Rome" to dependence on an individual commander, Chinese youth moving toward tang ping and bai lan are quietly disengaging from the narratives of "struggle" and the "great rejuvenation of the Chinese nation." That disengagement erodes the foundation of the state's mobilization narrative — silently, but persistently.
There is the story of the Gracchi brothers. In the 130s–120s BCE, Tiberius and Gaius Gracchus attempted to redistribute public land seized by the great landlords back to the smallholders. The entrenched senatorial aristocracy resisted. Both brothers met violent political ends. Structural reform failed. The trajectory of smallholder decline did not change.
In 2021, Xi Jinping declared "common prosperity" (gongtong fuyu, 共同富裕) — the idea of shared affluence. Regulations targeting Big Tech followed, along with heavier taxation on the wealthy and a crackdown on the private tutoring industry. Alibaba's Jack Ma (马云) vanished from public view before quietly resurfacing. The resistance of the entrenched Big Tech founders was not overt, but it was effective. As the common prosperity campaign faded, so did the regulatory pressure. The job market did not improve. Property prices continued to fall.
Both collided with entrenched interests. In neither case did the effort reach fundamental structural reform. The momentum generated by a system's own success — the force that displaced the very people who built that success — could not be stopped.
The achievement called "the world's factory" produced the automation and AI that are now displacing the workers who powered that factory. The structure that repeated in Rome is repeating in China. Only the era and the setting are different.
Transition: Four Walls Tightening on Each Other
The age-35 crisis. Tang ping and bai lan. Negative-equity real estate. The demographic cliff. These four are not independent problems.
The age-35 crisis feeds tang ping. Tang ping delays childbirth, accelerating the demographic cliff. Negative equity kills consumption, which drags down the economy, freezes the job market, and strengthens tang ping. Each wall reinforces the others. In China, these problems are unfolding at greater scale, at faster speed, with the new layer of AI shock stacked on top.
Wang rides the streets of Hangzhou again today. A notification arrives. 1.2 kilometers. Twelve minutes. The Python tutorial continues inside his AirPods. The structure has made it a rational choice for a computer science master's graduate to deliver food.
But on the other side of that structure, a different kind of person exists. People who, under the same sanctions and the same demographic cliff, stunned the world with scarce GPUs. People for whom constrained resources forced algorithmic innovation.
In the same space where China's displaced live, what are China's discerning building? If DeepSeek (深度求索) is part of the answer, what does the full picture look like?
Next chapter — Ch. 11: China's Discerning — Innovators Under Sanctions