10% of U.S. households own 88% of the country's total wealth. How will the pie be divided in the AI era?
From 2019 to the first quarter of 2026, the equity wealth directly held by U.S. households rose from approximately $29 trillion to around $55 trillion, nearly doubling. The rapid expansion phase was concentrated after 2023, which is the period of stock price rise driven by AI
AI (Artificial Intelligence) is gradually exerting its impact on income distribution.
On July 27, the *Quarterly Macroeconomic Policy Report (Q2 2026)* released by the China Finance 40 Forum (CF40) shows that in the past few years, AI technology has accelerated the expansion of the wealth pie, while also exacerbating inequality in income distribution.
From 2019 to the first quarter of 2026, the equity wealth directly held by U.S. households rose from approximately $29 trillion to around $55 trillion, nearly doubling. The rapid expansion phase was concentrated after 2023, which is the period of stock price rise driven by AI.
The report further analyzes that from 2022 to the first quarter of 2026, the equity wealth directly held by U.S. households increased by about $21 trillion, of which the top 10% of households captured approximately 88% ($18.5 trillion) of the total wealth, while the bottom 50% of households only received around 1% ($0.2 trillion).
The changes in U.S. household wealth initially reflect the income distribution imbalance brought by AI technology: wealthy households have taken a larger share of the pie.
At the press conference for the aforementioned report, Huang Yiping, a CF40 member and Dean of the National School of Development at Peking University, shared academic research findings showing that every major technological advancement in history has had a far-reaching impact on the income distribution structure.
According to research by economic historian Robert Allen, the so-called "Engels' Pause" emerged during the First Industrial Revolution, which means that in the decades after the outbreak of the First Industrial Revolution, the per capita output of workers increased significantly, but the real wages of workers did not rise notably. From the perspective of distribution proportion, the share of labor income in the total economic output did not begin to rise until about 100 years after the First Industrial Revolution broke out.
This also raises a question: Will there be a period similar to the "Engels' Pause" characterized by "stagnant wages, rising capital returns, and widening inequality" in the early stage of AI technological progress?
Huang Yiping believes that AI, as a general-purpose technology, is a historic opportunity for China's economic development. At the same time, AI technological progress may affect income distribution through four major mechanisms, leading to further imbalance in the income distribution structure in the coming period.
"We should attach great importance to the issue of income distribution and take precautions before problems arise," Huang Yiping said. Income distribution is closely related to total social demand. If the current trend continues, the pattern of strong supply and weak demand may be difficult to reverse in the short term, and is very likely to further intensify in the future, thus undermining the sustainability of economic growth.
Accordingly, Huang Yiping suggests adhering to the "people-oriented" principle, building a systematic policy package centered on the "Invest in People" strategy, balancing short-term social stability and long-term economic efficiency, and realizing inclusive sharing of technological dividends.
Four Major Mechanisms Affecting Income Distribution
While AI has set off successive waves of frenzy in the capital market, the haze of layoffs and job replacement continues to hang over working people.
The CF40 report shows that after the emergence of AI, the situation where capital expenditure replaces labor employment is actually happening.
U.S. market data shows that from 2024 to 2025, the employment growth rate in industries with high AI application penetration such as the information industry, professional services, finance and insurance was significantly lower than the average growth rate from 2010 to 2019. At the same time, AI has driven a substantial increase in the value of the U.S. stock market and a very uneven growth of household wealth: the top 10% of households captured about 88% of the total wealth, while the bottom 50% only got around 1%.
As a result, differentiation has emerged in the disposable income of residents: in 2025, the share of the top 10% of households in total disposable income rose to 36%, significantly higher than the average of 34.3% during 2010-2019; the share of the bottom 50% of low- and middle-income households in disposable income was 11.5%, a slight decrease from the average of 11.8% during 2010-2019.
Dramatic changes in income distribution driven by technological progress are not unprecedented. Robert Allen's research shows that the First Industrial Revolution started in the 1760s, but the share of labor wages in economic output continued to decline and did not stop falling and rebound until the 1870s. In other words, about 100 years after the First Industrial Revolution brought a sharp rise in human productivity, the share of the pie allocated to workers began to expand.
Huang Yiping said that every industrial revolution in history has had a profound impact on income distribution. The First Industrial Revolution brought the "Engels' Pause" that exacerbated inequality, where capitalists took a larger share of wealth. The Second Industrial Revolution gave investment in intangible assets such as technical knowledge and organizational capital an advantage in distribution. The Third Industrial Revolution saw employment polarization and wage inequality, that is, the income and employment opportunities for high-paying and high-skilled groups grew rapidly, low-end jobs remained relatively stable, while middle-tier jobs shrank and wages for medium-skilled workers stagnated.
As a widely recognized general-purpose technology, AI will undoubtedly bring another leap in human productivity. This time, how will technological progress act on income distribution?
Huang Yiping said that AI technological progress may affect income distribution through four mechanisms.
The first is the capital-biased mechanism, which is reflected in the continuous decline of the labor income share. Enterprises continuously increase capital input to replace traditional labor factors, thus changing the allocation logic of production factors and the income distribution pattern. In other words, AI empowerment can increase the per capita output of workers, but the income of workers does not grow accordingly.
The second is the task replacement mechanism, whose manifestation is similar to the hollowing out of the middle class and the "K-shaped" differentiation that emerged during the Third Industrial Revolution.
"After AI technology is implemented, will you be replaced or empowered? If you are empowered by AI, you will have more opportunities in the future, and your income will rise along the upward branch of the 'K' shape; if you are easily replaced by AI technology, your future income or returns are likely to go down," Huang Yiping said.
The third is skill differentiation and the digital divide, which may hinder class mobility. The technological gap between different groups is directly translated into income gap and unequal development opportunities, leading to industry and regional differentiation and a winner-takes-all effect. "In short, either you have a platform or you have the required skills, otherwise the development of new technologies may not be particularly favorable to you."
The fourth is the wealth distribution amplification mechanism, which is mainly reflected in the decoupling of capital returns from labor returns. Specifically, through asset appreciation, cost shifting and intergenerational resource transfer, wealth inequality is amplified on the capital side, forming a situation where "the rich get richer". "This is a common phenomenon and not strongly correlated with AI," Huang Yiping added.
Breaking the Monopoly of AI Dividends
Does the improvement of productivity necessarily bring faster economic growth?
Zhang Bin, lead author of the aforementioned report, senior fellow of CF40 and Deputy Director of the Institute of World Economics and Politics of the Chinese Academy of Social Sciences, said that the answer may be no, because AI may lead to slower growth on the demand side than in the past, and the economic growth rate is usually determined by the weaker side between supply and demand.
The reason why the impact of AI on income distribution has attracted so much attention is not only that it concerns the fate of individuals and households, but also that it may exert a far-reaching impact on total social demand.
"Working groups mainly rely on salary income, and their marginal propensity to consume is significantly higher than that of capital owners who rely on asset appreciation. When AI transfers wealth from laborers to capital owners, the whole society will face a serious shortage of total consumer demand," the CF40 report cites an academic paper published in 2020.
In recent years, the feature of "strong supply and weak demand" has persisted in China's economy.
In the first half of 2026, China's real GDP (Gross Domestic Product) growth reached 4.7%. Among them, the total retail sales of consumer goods increased by 1.3% year on year, and the national fixed asset investment (excluding rural households) decreased by 5.7% year on year, indicating that domestic demand is generally weak.
Huang Yiping said that the driving effect of AI on the supply side is already very obvious, but if total demand does not pick up and supply is too strong, the economy will not be sustainable. "If the current trend continues, it is very likely that a scenario will emerge in the future: the pattern of strong supply and weak demand cannot be reversed in the short term, and may even further intensify," Huang Yiping said. "We need to consider how to continuously increase total demand to keep it relatively balanced with supply."
In Huang Yiping's view, AI is a historic opportunity for China's economic development, and at the same time, we should attach great importance to the issue of income distribution and take precautions to cope with the possibility of further imbalance in income distribution in the future.
In this regard, Huang Yiping put forward three coping strategies.
The first is defense: by consolidating the bottom line of social security, we can buffer the unemployment impact brought by AI, while curbing the excess monopoly gains formed by capital with the help of algorithms, so as to safeguard the basic foundation of social fairness.
Specific measures include improving the unemployment insurance and safety net system; strengthening anti-monopoly law enforcement in the platform economy and AI sectors; establishing ethical review and restriction mechanisms for purely replacement-oriented AI applications.
"In the short term, it is certain that AI technological innovation will have a great impact on some occupations, and some people will lose their jobs. Every industrial revolution in the past has produced such an effect. This is a normal phenomenon, but the key issue is how to achieve a smooth transition," Huang Yiping suggested that we should adhere to the "employment first principle", encourage AI technological innovation that empowers labor rather than replaces labor, and at the same time ensure that no major social problems arise by consolidating social security safeguards.
The second is empowerment: reshape the human capital structure, promote workers to shift from "being replaced by AI" to "mastering AI innovation", and realize the collaborative symbiosis between humans and intelligent technology through capability upgrading.
Specific measures include reforming the education system to integrate AI literacy and creative thinking training; establishing a lifelong learning vocational skill training system; promoting skill certification and employment support for "AI + Occupation" programs.
"In the future, the relative importance of school diplomas for young people to find jobs and seek career development may decline, while the importance of composite skills will rise," Huang Yiping said. The "Invest in People" strategy proposed by our government has a very important aspect, which is to cultivate people's ability to collaborate with AI.
The third is rebalancing: restructure the distribution mechanism of production factors, break the monopoly of capital and technology over AI dividends, and ensure that the value created by data and algorithms benefits a broader group of social members through institutional design.
Specific measures include exploring the imposition of adjustment taxes on excess AI returns; clarifying the ownership of data property rights and promoting the socialized sharing of public data; establishing an AI dividend distribution fund shared by all people.
"How to make the whole society share the value created by technology and algorithms puts forward very high requirements for public policies," Huang Yiping admitted. It may be too early to consider a universal AI dividend distribution fund for all people at present, but we may consider providing some living or income support for low-income groups.
"Judging from the current situation, the problem of overall income distribution inequality is quite prominent, and it may become even more prominent in the future, while insufficient demand is the most prominent problem we are facing at present," Huang Yiping said. Turning the "Invest in People" strategy into a complete set of policy plans may help alleviate the current problem of strong supply and weak demand and improve income distribution.
(The author is a reporter of *Caijing*)
This article is from the WeChat Official Account "Caijing May Flower" (ID: Caijing-MayFlower), written by Tang Jun, edited by Zhang Wei, and authorized for release by 36Kr.