Skip to content
On the 250th anniversary of The Wealth of Nations, a new argument holds that technology creates lasting monopolies Adam Smith never anticipated, and that AI could deepen inequality. (Illustration)

By Mordecai Kurz, Project Syndicate | November 22, 2026

For all that he got right, the "founding father" of economics failed to see how technological innovations lead to market monopolization. If AI is left to the invisible hand, it will almost certainly increase economic and political inequality, causing even deeper social polarization and further undermining democracy.

STANFORD—This year marked the 250th anniversary of the publication of Adam Smith’s The Wealth of Nations, a book that remains widely—and deservedly—celebrated. But this moment in economic history demands that we address one of Smith’s gravest and most consequential errors.

Having witnessed the early stages of the Industrial Revolution, Smith recognized the importance of technological innovation for economic progress, as well as the harm that monopolists cause. But he thought that monopoly power came only from royal charters and mercantilist policy. He failed to see that, under the free-market policy he advocated, technological innovation also creates monopolies and market power. Where a firm dominates a market, Smith’s famous “invisible hand” fails, leading to permanently inferior economic performance.

Although a free-market policy effectively promotes economic progress, the market conditions it creates can hardly be described as “free.” Innovators secure market power (what I call “the market power of technology”) through patents and trade secrets, which ultimately confer monopoly profits. We can see this clearly in the United States, where presidents from both parties have upheld the free-market orthodoxy that Ronald Reagan embraced in 1981. With more than four decades of evidence, one can study the real-world, as I did in my past two books, The Market Power of Technology (2023) and Private Power and Democracy’s Decline (2026).

The question of how—or whether—to regulate AI has brought this issue to the fore once again, with those arguing for a hands-off approach to the technology basing their views on ideas dating back to Smith. Fully understanding Smith’s error is thus a matter of political urgency.

The post-1981 free-market orthodoxy’s treatment of technological innovation has had two important consequences. First, innovations cause economic inequality to soar, with monopoly profits permanently accounting for as much as 25% of corporate income. Second, the same innovations unleash forces that undermine democracy and weaken its institutions. For at least a century, Americans have been told that economic freedom and political freedom are two sides of the same coin. Yet nothing could be further from the truth (though some leading scholars have stuck to their guns).

Given that free-market capitalism itself threatens democracy, the contemporary debate over whether AI will destroy humanity or usher in a golden age of abundance is too narrow. We surely do need safeguards for how this technology is developed and used, but we also need to regulate how its benefits and costs are distributed. To avoid repeating the experience since the 1980s, we must establish a legal obligation to restore the livelihoods of all workers whom the technology supplants, so that nobody is left behind. Unless we establish policies that address the market power of technology, we can confidently predict that AI will increase economic and political inequality, cause even deeper social polarization, and further undermine our democracy.

First Mouse Gets the Cheese

Economists have long assumed that market power would be short-lived: competition would erode margins, patents would expire, and trade secrets would not last. Allowing innovators to amass short-lived market power has been seen as a small price to pay for the great benefits of rising productivity. But that assumption is no longer viable. As I documented in an earlier commentary, the evidence since 1981 shows that innovators’ monopoly power is neither short-lived nor a small price to pay.

Some innovations develop in isolation, while others spark a competitive race. Either way, only one or two winners emerge, namely the innovators with first-mover advantages. It is they who set industry standards and build consumer confidence, even if later products are superior, as with VHS. They gather information about consumers and suppliers, achieve operational scale that lowers incremental costs, secure permits and access to financing, and choose an optimal location.

Without additional strategies, the first-mover advantage would indeed dissipate as patents expire and trade secrets are revealed. Yet innovators are well aware of this risk and use diverse strategies to consolidate their market power. One common strategy is to deploy regular technology updates, thereby creating a patent pyramid: an interdependent combination of old and new patents that effectively extends a patent’s life beyond its intended span.

A second, lethal strategy is acquiring potential competitors, their technologies, or their leaders (through so-called acqui-hires). Firms may then use the acquired technologies or suppress them to protect their own technologies from rapid obsolescence. Microsoft acquired 237 firms from 1987 to 2020, and Google acquired 236 from 2001 to 2020, as a means of establishing technological empires that incorporate products others had innovated.

A third strategy to maintain market power is to suppress potential competitors through various hostile acts aimed at impeding their development, as when Microsoft restricted its distribution of Netscape’s web browser. The company also used a fourth strategy with its Windows operating system: build an interdependent ecosystem around a unique proprietary operating system, then introduce products and devices linked to the core technology. This “tying and bundling” approach has the virtue of expanding one’s market because customers buying one product must also purchase others on the platform.

A fifth strategy is building proprietary consumer and supplier information banks, a key advantage for advancing AI. Doing so has helped Microsoft, Google, and Amazon dominate cloud services, where data processing is now concentrated. Lastly, companies can lock in their positions with loyalty programs, which are now universal.

In addition to the market power of their technology, information-technology firms also benefit from economies of scale and network externalities that reduce incremental costs. The larger the scale of the established innovator, the harder it is for new entrants to compete.

In 1956, the economist Joe S. Bain observed that no entry occurs in markets where firms earn monopoly profits, even without a technological advantage. Building on this empirical insight, game-theoretic research has produced many papers showing that an incumbent monopolist has many commitment strategies to deter new entrants. There is now a vast literature showing that firms winning the initial technological race retain monopoly power for a long time.

The Limits of Competition

Why does competition fail to erode this market power? In regular price competition, firms earn monopoly profits when price exceeds incremental cost, and incur losses when it falls below. They adjust output simply by changing their inputs until price equals incremental cost. In the end, all firms survive and earn no monopoly profits.

Technological competition is very different. It begins with an incumbent who owns the technology. Rivals need more than just inputs to production; they also must invent somethingbetter—a formidable task—and even then, the incumbent has the advantage of being able to use any of the strategies listed above. The usual outcome is that one technology wins, driving competitors out and leaving a monopolist in place. In the rare case when an innovator displaces an incumbent monopolist, a more powerful monopolist emerges.

Many view Joseph Schumpeter’s theory of “creative destruction” as a consistent model of technological competition. In its original form, it makes the common-sense observation that new technologies make older technologies obsolete and eliminate their market power. But this is too simplistic. Firms die for various reasons (think of Sears, Xerox, Polaroid, and Enron) and new firms are born (the likes of Apple, Microsoft, and Amazon). Some fail simply because their management makes the wrong decisions (think of Nokia’s response to the smartphone).

But poor management decisions are not relevant to why the market power of technology is permanent. To understand that, we must consider the technologies involved.

We can start by distinguishing between innovations within a technological paradigm and those that transform the paradigm. Paradigm-changing innovation reshapes human knowledge, drives societal evolution, and enables applications across the economy. These rare technologies are discovered roughly once every half-century and are called general-purpose technologies (GPT). Examples include the steam engine, electricity, the internet, and now AI.

New GPTs do create competition. Older firms lose their market power and must reinvent themselves through research and development, joint ventures, or acquisitions. In the race to apply the GPT across all sectors, old and new firms are on equal footing. Some older firms may retain their lead, some younger firms will fail, and others will rise to leadership. Thus, technological competition unfolds as expected, but only in response to a new GPT.

The evidence shows, however, that most technological improvements come from incumbent firms, not challengers. Many big firms today—including DuPont, Johnson & Johnson, Procter & Gamble, and General Electric—are over a century old, with some dating to the 19th century. All have undergone repeated technological renewal.

Crucially, the market power of technology is legal under US antitrust law. For example, Apple does not monopolize smartphones. Consumers can choose other brands, and competitors are free to enter the market. Nonetheless, Apple controls its technology, which others cannot use. To challenge Apple, a competitor must invent a new technology that consumers consider superior to the iPhone—a difficult task. That is why Apple can charge such high margins, capturing 44% of global smartphone revenue in 2021 while accounting for just 15% of global smartphone sales.

A final reason for limited technological competition is firms’ preference for cooperation, which is legal. Incumbent monopolists use past profits as a “war chest” to deter challengers by enticing them. Why wage a long, costly war against an incumbent who can strengthen its position in many ways, especially when being acquired may be more rewarding? A successful, valuable startup must choose between becoming wealthy overnight or risking everything for a chance to earn even more. Most choose acquisition.

Even large firms prefer cooperation over conflict. General Electric, founded in 1892, acquired the patents for the first DC (direct-current) electric bulb and power generation technologies. In 1886–87, Nikola Tesla patented the AC (alternating-current) motor and sold the rights to Westinghouse. This sparked the “war of the currents,” which AC won. Rather than eliminate GE, Westinghouse reached a patent-sharing agreement with it, and the companies jointly monopolized the market for the next half-century.

Silicon Valley startups have similar motives. Rather than replace Google or Microsoft, almost all aim to establish a proof of concept and be acquired by a leading firm. This preference is reflected in the changing number of US public firms. Non-financial firms with publicly traded stocks in the Compustat files rose from 3,914 in 1980 to 7,429 in 1998 during the IT revolution, then fell to 4,621 by 2016. Most of the 2,808 missing firms did not fail; they were acquired, as were many privately-held firms.

The AI industry today offers further evidence of such cooperation. Microsoft turned potential competition with OpenAI into a $13 billion partnership. Microsoft, Google, and Amazon are all heavily invested in Anthropic, and Nvidia backs nearly every AI firm. Google, Meta, OpenAI, Anthropic, Apple, and Samsung have multiyear partnerships with Broadcom to develop custom AI accelerators (XPUs). Similarly, in pharmaceuticals, smaller firms do most of the research, aiming to demonstrate clinical benefits and to be acquired by one of the larger companies.

The Costs of Rising Market Power

Firms that gain market power earn monopoly profits, and the rising value of their securities reflects their increased monopoly wealth. The empirical evidence for this is clear. Today’s free-market economies revolve around massive firms, each dominating one or several market segments, where it generates extremely high monopoly profits. Monopoly wealth is the market capitalization of those profits.

The magnitude of monopoly profits and monopoly wealth is immense. I estimate that in 1980, monopoly profits and monopoly wealth were negligible. Yet by 2019, monopoly profits accounted for about 24% of total income created by the US corporate sector. Meanwhile, monopoly wealth rose by $25.1 trillion from 1980 to 2017—accounting for over half the total value of stock traded in 2017—and likely exceeded $35 trillion in 2026.

Most of those gains went to a small group of people: innovators, financial advisers, and venture capitalists who buy shares cheaply before a company goes public. When innovation succeeds, a stock’s value surges, rapidly enriching the early investors. This is the only way to amass vast wealth in one lifetime, and it explains how 73% of America’s 989 billionaires (as of 2026) earned their fortunes.

Of course, innovation and technical improvement have always left some groups behind. The digital technologies developed since the 1980s have nullified the gains of unskilled workers without a college education, who benefited from the moving assembly line invented in 1913. Whereas the assembly line created blue-collar jobs that paid middle-class incomes, computers and robots have caused a sustained decline in blue-collar jobs. That is why the real (inflation-adjusted) wages of low-skilled workers, who represent the majority of the US labor force, barely changed between 1980 and 2021.

At the same time, these forces benefited college graduates and drove the incomes of tech entrepreneurs, corporate CEOs, and some professionals sky-high, worsening inequality. Political leaders largely ignored the massive job losses experienced by a large segment of American workers, and these workers lost faith in the American political system—one of the central reasons for democratic backsliding.

The wealthy have many opportunities—from local school boards to the presidency—to exert political influence. Elon Musk is merely the most vulgar example of a much larger problem. US Supreme Court rulings on campaign finance have made it easier for the wealthy to gain political power. Since the 1980s, the influence of money, and of ultra-rich individuals like Musk, Peter Thiel, and Miriam Adelson, has grown sharply.

The rising market power of technology has created centers of private influence built around leading firms, their officers, and major stockholders. Acting either as organizations or as wealthy individuals, they can and do wield immense resources to advance their interests. They influence legislation through massive lobbying and campaign funding, electing friendly politicians and defeating those who oppose them. The evidence also shows that rising wealth inequality has convinced ordinary citizens that democratic institutions do not work for them. Rising private power has eroded democracy’s legitimacy.

So far, the evidence on AI’s effect on inequality is deeply concerning. Income and capital gains generated by several AI firms, such as Nvidia, Broadcom, OpenAI, and Anthropic, already show extreme concentration, creating a growing number of billionaires. All indications point to AI enriching technologists and owners of technology while impoverishing most workers. This suggests that, left to the market, AI’s effect on income and wealth inequality will be dramatic.

The Right Reforms

What is to be done? In the US, the most urgently needed political reforms include reducing the role of money in politics, outlawing gerrymandering, eliminating the Electoral College for presidential elections, and removing the president’s absolute immunity from prosecution while in office.

To reconcile economic efficiency and rising productivity with democracy, capitalism must be regulated. Economic policy must actively reduce market power and extreme economic inequality, and AI must be developed so humans retain ultimate control. Its design should aim to support human labor rather than replace it, and regulators must limit its impact on inequality to safeguard democracy. Therefore, economic reforms should reflect three principles.

First, market power must be contained. That means restructuring patent law, limiting mergers and acquisitions, raising taxes on corporate profits and the top personal income tax bracket, and eliminating restrictions on collective bargaining.

Second, the economic gains from technology must be shared more equitably. This can be done by establishing guardrails on AI that favor innovations that complement human labor rather than replace it; upskilling service jobs to boost productivity and pay; raising the federal minimum wage; and embracing the social, morally uplifting role of labor unions.

Third, preserving citizens’ livelihoods must be a top priority.To that end, I have proposed a federal legal obligation to restore the earning capacity of any worker whose job is eliminated by a publicly supported measure. Displaced workers would be entitled to full financial support during retraining, as well as help finding a new job. Scandinavia, Germany, and Japan developed successful policies along these lines, helping to stabilize their democratic institutions.

Power-worshipping oligarchs dominated America during the 19th-century Gilded Age, but reform-minded movements and leaders emerged to break their political grip. The period from Theodore Roosevelt and the Progressive Era to Franklin Roosevelt’s New Deal restored democratic stability for almost a century. Now that a half-century of free-market policy has brought oligarchy back, our generation faces a similar task.

Mordecai Kurz, Emeritus Professor of Economics at Stanford University, is the author of The Market Power of Technology: Understanding the Second Gilded Age (Columbia University Press, 2023) and Private Power and Democracy’s Decline: How to Make Capitalism Support Democracy (MIT Press, 2026).

💬 Worth Your Time?
Loved it, hated it, or somewhere in between — tell us. Grade this article here and help shape what lands in your inbox next.

Comments

Latest