California may put an extraordinary wealth tax before voters this fall. Under the proposal, anyone worth more than a billion dollars would see 5 percent of their wealth taken next year. Supporters cast it as a one-time measure targeting only the super-wealthy, yet that reassuring description may mislead voters. A state that enacts such a tax to fill its empty treasury will be tempted to do so again and even expand it when those coffers run dry. Meanwhile, California is not alone. Massachusetts Senator Elizabeth Warren has proposed a yearly federal wealth tax of 2 percent on fortunes above $50 million and 3 percent on those above $1 billion.
Friends of liberty should be worried by this novel means of raising money in the United States, even apart from the significant questions about their constitutionality. Wealth taxes fall even on productive, illiquid capital as if it were idle treasure rather than the dynamo of our economy. They would require a vast administrative apparatus to collect, while generating endless forms of legal tax avoidance. Today, by threatening the capital of wealthy innovators, wealth taxes also may endanger national security and even our longevity. AI is a public good essential to our military, as recent coverage in the New York Times acknowledges. Its knowledge spillovers also accelerate scientific discovery. And progress in AI benefits greatly from founders’ enterprises and venture capital, whose advantageous deployment will be most discouraged by such taxes. In short, wealth taxes are terrible not because redistribution is necessarily always wrong, but because they tax ownership rather than realized gains and fall especially hard on entrepreneurial and strategically important capital.
Classical liberal and commercial republics have generally not taxed ownership of productive assets. The economic growth they create helps maintain social peace. As the political philosopher Ernest Gellner observed, modern democracies rely on “sustained and perpetual growth” to avoid the dissension that comes from groups fighting over a static or shrinking pie. But the costs are not only economic. As I describe in my recent book, Why Democracy Needs the Rich, a free society depends not only on limits on government, but on independent centers of initiative and judgment outside the state. A tax on ownership weakens those centers and increases political control over all aspects of our lives.
Wealth taxes in America have traditionally targeted only land or housing. These assets, which are visible and easier to value than private businesses, are not sources of innovation. Because of these qualities, they do not substantially harm productivity or impose high administrative costs on collection.
To be sure, the United States and other advanced industrial nations often, though not uniformly, tax capital gains when realized, unlike ordinary income, which is taxed as it is earned. The realization requirement, however, prevents illiquid capital from being valued and taxed. This provision does not require founders of companies to sell their shares when their control is most needed. Even so, taxes on capital income are generally lower than those on ordinary income, a distinction that reflects the socially beneficial productivity of capital.
One way to illustrate the harm of the wealth tax is to translate its costs into what they would be as an effective income tax. Assuming a five percent inflation-adjusted return on capital, a five percent wealth tax wipes out all income. Even a 2 percent tax on capital equates to a 40 percent tax on capital gains at that rate of return. Add the 23.8 percent capital gains tax, and we are at 63 percent. California imposes another 13 percent tax for its top rate, bringing the total to 76 percent. Moreover, because a wealth tax is due regardless of whether an asset produces cash, it can force sales in bad years, precisely when retaining ownership helps avoid losses from cyclical swings. This is hardly a recipe for encouraging productivity.
Wealth taxes impose a significant administrative burden. Unless an asset is publicly traded in a liquid market, there is no objective valuation. Collectibles, for example, may be hidden in homes. Administering a wealth tax thus creates a dilemma: the government must either greatly expand its tax-collection force or accept substantial underpayment. In either case, the result is capital flight and more creative legal tax avoidance. Again, this is not just an economic problem, but a political one. When valuation is disputable, taxation loses clear rules and gives officials arbitrary discretion.
AI shows that the ultrawealthy can accelerate progress that yields cheaper services, better medicine, and improved education for everyone.
Because of such problems, European nations that once imposed wealth taxes have largely abandoned them. The Organization for Economic Cooperation and Development (OECD) has noted that wealth taxes create more economic distortions than other forms of taxation. Moreover, they are less fair, because they apply regardless of the actual return on capital, unlike taxes based on realized income. As the OECD observes in the abovementioned report, there are fairer and more effective ways to redistribute money and reduce wealth inequality if that is considered a useful social goal. Progressive taxes on income from capital (which the United States already has) and inheritance taxes, for example, can directly address wealth inequality. In my view, eliminating the tax-free capital gains reset at death that heirs enjoy, even on large estates, would be another step the United States could take to make its tax system more progressive for the very wealthy without the distortions of a wealth tax. If the concern is entrenched inequality driven by inherited wealth, taxing transfers at death is far less damaging to economic activity than taxing capital deployment during life. And even those who think some great fortunes reflect privilege rather than merit should resist a wealth tax. The better remedy is to attack monopoly and rent-seeking directly (often by imposing more limits on government power to prevent market entry), not to tax productive ownership.
A wealth tax pushes investors away from risky but socially useful bets and toward liquidity and short horizons. Thus, wealth taxes in general, and California’s in particular, are dreadfully timed because they would weaken both the incentives for entrepreneurship and the supply of capital needed to advance AI. That use of capital should remind us that wealth is about more than funding luxury consumption. It can be, and often is, the source of public good development. AI is the most important public good in generations, because it is a general-purpose technology with hugely positive knowledge spillovers. More powerful AI also accelerates scientific research. DeepMind’s AlphaFold has revolutionized protein analysis, making two hundred million protein structures available and enabling biotech researchers to design vaccines and discover new antibiotics more effectively. That same breakthrough is helping researchers discover materials for new batteries and solar panels to help replace fossil fuels.
Technology has always shaped military success, and in our pervasively high-tech age, it is becoming even more decisive. For instance, AI processes large volumes of data from satellites and other sources, enabling drones and other defense systems to operate even in environments where communications are disrupted. If the US does not develop such capabilities, we can be sure that our adversaries, including authoritarian nations like China and Russia, will.
Ultrawealthy individuals, either directly or through technology firms they continue to lead, have coordinated the AI models driving this progress. OpenAI, founded by Sam Altman, pioneered the generative chatbot—transforming it into the fastest-growing consumer technology ever. Other labs have, at times, surpassed OpenAI in capabilities; for example, Anthropic, led by Dario Amodei. Google has also played a significant role, having acquired DeepMind and continuing under the leadership of Nobel Prize Laureate Demis Hassabis. By some measures, Google had the most powerful model at the end of this year. Meanwhile, Elon Musk has launched his own AI company.
AI cannot advance as effectively without the support of current and future entrepreneurs. The government can fund basic science but lacks the concentrated technical talent and research culture to produce this public good. Most importantly, it lacks the knowledge that, as Friedrich Hayek argued, emerges only from the trial-and-error of competition. Decentralized capital allocation is essential for that discovery, but wealth taxes interfere with it. Founders’ vision is also part of that discovery process, but wealth taxes undermine that control at the fragile stages of firm growth. Venture capital makes bets on these founders, but wealth taxes discourage these risky investments. This concern is not speculative: American venture capital and investors poured $109 billion into AI in 2024. Some of that comes from the virtuous Silicon Valley circle in which successful AI entrepreneurs reinvest their profits in promising new companies.
The system also encourages brilliant graduates working in garages who can accelerate the process. Venture capital provides the resources and guidance for unconventional approaches that may prove successful. These are precisely the advantages that the United States enjoys over China. Authoritarian regimes can mandate AI development, but they lack the deep, decentralized discovery process that these rich entrepreneurs and investors create.
Politicians decry the inequality generated by tech billionaires. But AI also shows that the ultrawealthy can accelerate progress that yields cheaper services, better medicine, and improved education for everyone. Of course, their social utility does not mean we should not have more progressive taxes. That is a complicated question of getting the incentives right. But the fact that the United States is by far the AI leader in the free world suggests that our incentives for cultivating the culture that produces this public good are not poorly calibrated. The rest of the West here, as elsewhere, free rides on our achievements.
The ultrawealthy have always been a target of envy—and like other sins, envy distorts reasoning. It may be that even after reforming wasteful government spending and entitlements, we will still want to raise some taxes on the wealthy to address our fiscal imbalances. But wealth taxes are an especially destructive way to do that. In an AI age, they are not merely bad tax policy. They are downright dangerous.
