Americans’ retirement savings are disproportionately tied to the dozen Big Tech firms that now dominate the S&P. This makes any intervention into regulating Big Tech that risks devaluing them politically difficult, writes Hera Hyeonseo Lee.
Here is a number that should unsettle anyone with a retirement account: the ten largest companies in the S&P 500 now account for roughly 40% of the entire index’s weight. Nearly all of them are AI-adjacent tech firms: Nvidia, Microsoft, Apple, Amazon, Alphabet, Meta, Broadcom. When you contribute to your 401(k) and your money flows into a target-date fund tracking the S&P 500, about forty cents of every dollar goes to these companies. You almost certainly did not choose this. The architecture chose it for you.
The dominant debate about AI asks whether the industry is a speculative bubble. The cyclically adjusted price-to-earnings ratio for the U.S. market has exceeded 40 for only the second time in history, though some analysts argue current valuations are supported by real earnings growth. This framing misses a prior question: even if a correction is warranted, is it politically achievable? The financial architecture of American retirement has created a structural constituency for AI capital—a mass base of involuntary shareholders whose material interests are now aligned with the continued growth of a handful of tech firms, whether or not those firms deliver on their promises. This constituency does not need to be organized. It is architected.
The default architecture
The mechanism begins with how Americans save for retirement. Since the Pension Protection Act of 2006, employers have been encouraged to auto-enroll workers into 401(k) plans and designate target-date funds—diversified portfolios that automatically shift from stocks to bonds as the worker ages—as the default investment. By 2024, 61% of Vanguard-administered plans had adopted auto-enrollment, achieving a 94% participation rate among those plans. Eighty-four percent of participants held assets in target-date funds, and 71% of target-date investors held their entire account in a single such fund.
Auto-enrollment is widely celebrated as a triumph of behavioral economics—the insight, advanced by Richard Thaler and Cass Sunstein, that people save more when the default is to participate. The policy has been effective at getting workers into retirement plans. What is less discussed is what they are saving into. Target-date funds are designed to be left alone. The fund rebalances automatically. The result is that the vast majority of retirement savers passively channel contributions into whatever the market-cap-weighted index happens to contain.
And what the index now contains is an extraordinary concentration of AI capital. The top ten stocks represented about 19% of the S&P 500 in 2015. By 2025, that figure had more than doubled to nearly 41%, while those companies accounted for only 32% of the index’s earnings. A municipal employee in Texas who was auto-enrolled in 2018 and never adjusted her allocation now has a substantial portion of her retirement riding on whether Nvidia’s data center revenue keeps growing. She did not choose this exposure. The default chose it for her.
The anti-antitrust constituency
This arrangement has direct implications for competition policy. Consider what happens when a serious regulatory intervention reduces the valuations of major AI firms: an antitrust action that breaks up Alphabet; a data privacy regulation that restricts the data pipeline used to train large language models; labor protections that impose costs on firms replacing workers with AI. Each of these policies would register, mechanically, as a decline in the retirement savings of the very workers such regulation is meant to protect.
No senator wants to explain to constituents why their retirement accounts dropped 15% because of a regulatory action the senator voted for. The political cost of intervening against AI market concentration is distributed across millions of households, while the benefits—less labor displacement, greater data privacy, reduced monopoly power—are abstract and long-term. This asymmetry creates a powerful bias against antitrust enforcement before any industry lobbying begins. AI firms do not need to spend a dollar on K Street to enjoy this protection. The default architecture of American retirement provides it for free.
The Stigler Center’s own research on digital platforms has documented how consumer welfare is shaped by market structure. The structural constituency I describe here suggests that market concentration in AI is defended by a financial architecture that aligns tens of millions of passive investors with the continuation of that concentration, whether they know it or not.
When wages fail, capital fills the gap
The trap has a second jaw. Labor’s share of national income in the nonfarm business sector stood at 54.4% in the fourth quarter of 2025, near the bottom of a range that has been falling for decades. When wages stagnate relative to the cost of living, retirement security increasingly depends on investment returns. Workers are not investing in AI firms because they believe in the technology. They are channeling savings into AI firms because their wages are insufficient to fund retirement through saving alone, and the default architecture routes their money into a market-cap-weighted index dominated by those firms.
The perversity is worth pausing over. The workers facing the greatest risk of displacement from AI automation—those in routine cognitive and administrative tasks—are disproportionately the workers who depend most on 401(k) returns to supplement stagnant wages. Their interest as retirement savers is directly opposed to their interest as workers. To the extent that AI valuations reflect expectations of labor cost reduction—and Wall Street analysts are explicit about this—these workers are financially invested in the acceleration of their own displacement. Some retirement savers may emerge from their AI investments with well-funded 401(k)s, but many workers will eventually have no or little income from which to save.
The foreclosure of correction
Market correction is partially inhibited because passive index flows are price-insensitive. Money flows into AI-dominant index funds based on contribution schedules, not fundamental valuation. Passive funds own roughly a quarter of the S&P 500 by market capitalization. These flows continue regardless of whether AI firms are overvalued. They are on autopilot.
Regulatory intervention is equally inhibited. Any policy that materially reduces AI valuations imposes immediate, visible costs on a broad constituency while delivering diffuse, long-term benefits. The political calculus is lopsided before any lobbying begins.
What would it take?
If this argument is right, then proposals to regulate AI market concentration—antitrust enforcement, data protection, labor transition programs—must be paired with proposals to restructure retirement finance. As long as workers’ retirement security is mediated by market-cap-weighted index funds routed through auto-enrollment defaults, their material interests as savers will conflict with their material interests as workers.
Resolving that conflict requires structural changes. One option is to decouple retirement from capital markets through expanded Social Security or a guaranteed retirement income. Another is to redesign index construction so that no single sector dominates the default. A third is to give workers genuine exit options from the funds they are automatically placed into. The Securities and Exchange Commission could require index funds to disclose concentration risk in terms savers actually understand. Fiduciary standards set by the 1974 Employee Retirement Income Security Act could be updated to consider whether extreme sectoral concentration in default funds serves participants’ long-term interests. Equal-weight index alternatives could be offered alongside cap-weighted defaults.
None of these solutions are simple. All of them are necessary before the regulatory correction that market skeptics and political economists anticipate can become politically viable. The architecture of American retirement is a political instrument, and AI capital is its primary beneficiary.
Author Disclosure: The author reports no conflicts of interest. You can read our disclosure policy here.
Articles represent the opinions of their writers, not necessarily those of the University of Chicago, the Booth School of Business, or its faculty.
Subscribe here for ProMarket’s weekly newsletter, Special Interest, to stay up to date on ProMarket’s coverage of the political economy and other content from the Stigler Center.
