Will the AI Economy Have a Middle Class? The Case for an AI Homestead Policy

Nevertheless, we should heed the warnings of the AI industry’s own leaders, such as Microsoft AI CEO Mustafa Suleyman’s expectations of the imminent transition for office workers and OpenAI CEO Sam Altman’s April 2026 call for redistributive robot taxes and a national wealth fund. These warnings may not go far enough. For many families, prolonged white-collar displacement would become more than an income problem. It would become a balance-sheet problem—forcing households to draw down savings, tap retirement accounts, delay homeownership, and interrupt the process through which wages become assets and assets compound over time. The domino effect would cut against the central promise of the American dream: that wage earners can become asset owners.

Several issues are vying for voters’ attention in the run-up to the 2026 midterm elections: affordability, the war with Iran, the data center buildout in their backyards, and next-generation fears about what AI will do to jobs. But they should demand a more fundamental debate about how to build a modern version of the wages-to-wealth ladder, akin to the one originally envisioned by Abraham Lincoln 170 years ago. In his 1859 address to the Wisconsin State Agricultural Society, he argued: “The prudent, penniless beginner labors for wages awhile, saves a surplus with which to buy tools or land for himself, then labors on his own account.” President Lincoln backed that vision with policy: The Homestead Act transferred 270 million acres of public land to numerous Americans, admittedly to the detriment of native peoples, on an unprecedented scale. The Morrill Land-Grant College Act put advanced education within reach of working-class families. The transcontinental railroad, financed by federal land grants, opened up markets and livelihoods across the country.

As technology transformed the economy, successive generations adapted the model. Progressive-era antitrust reforms opened industrial-age opportunities. The GI Bill and Veterans Affairs mortgages expanded homeownership after World War II, albeit with the exclusion of many Black veterans. The 401(k) created broad-based access to financial asset ownership beyond real estate. Through fits and starts, the United States built the rungs of a ladder connecting wages to wealth. The question today is whether policymakers can rebuild that ladder for an AI economy.

The current White House AI Action Plan, along with the accompanying Executive Order 14179, reflects an ambitious push for decisive American leadership in AI across science, exports, energy infrastructure, workforce, and innovation. It is fundamentally a bet that deregulation will deliver innovation with safety and productivity gains with new job creation. But that approach also risks accelerating the concentration of wealth and market power in an industry already dominated by a few large firms with extraordinary advantages in compute, capital, data, and distribution. And it underweights the possibility that market forces alone may not adequately account for the costs of the coming AI labor transition for workers—or the concerns of communities confronting large-scale AI infrastructure buildouts in their backyards.

The Biden-era approach was equally ambitious in its own way, relying more heavily on federal grants, subsidies, and rigorous safety standards to steer firms behavior, including where firms locate anchor investments and how innovation clusters form. But the risks of that approach were that different industrial policy mechanisms were often too slow and bureaucratic to meet the demands of the technology race. And they tended to overload incentives with social and environmental conditions that delayed or disincentivized the very investments policymakers were seeking to accelerate. Earlier industrial revolutions unfolded over decades. AI appears poised to reorganize labor markets, enterprise structures, and capital concentration within a few short years.

Neither approach focuses on the importance of preserving economic dynamism—the composite of individual capability, opportunity for firms and workers, and the connectivity of communities to broader markets—especially during technology transformations. The United States needs an AI Homestead policy—a reimagining of Lincoln’s concept. The original Homestead Act was not narrowly a redistributive tax on the gains of westward expansion. It gave broad swathes of the American population access to the nineteenth century’s productive frontier. An AI Homestead should represent a broad policy and private sector commitment to give Americans a fighting chance to own the AI economy’s productive frontier before those gains are locked in.

The AI Homestead starts with individual capability. The United States needs a massive retooling effort for workers of all ages as AI reshapes how work gets done across the economy. Incremental workforce programs will not be enough. Nor is it wise to leave it to individuals to figure it out purely on their own, given China’s massive AI diffusion efforts. While the government should not be in the business of providing training, Congress should establish AI training vouchers and targeted tax breaks to give workers an incentive to acquire AI-related skills. The goal is not simply to retrain workers after disruption occurs. It is to maintain the adaptive capacity of the American workforce before displacement hardens into economic exclusion. A growing body of evidence suggests the future of work may not be defined by humans competing against AI, but by AI-enabled workers pulling away from other workers without access to AI tools, training, and complementary skills.

Second is labor mobility. Roughly one-quarter of the workforce remains constrained by outdated occupational licensing barriers. Interstate credential portability and AI transition insurance—unemployment insurance for job losses directly related to automation—would make it easier for workers to move across sectors and regions as technology reshapes labor demand.

Third is business formation. Congress should extend Qualified Small Business Stock tax exclusions to small businesses deploying AI productively, not simply venture-backed startups building frontier models. Federally supported ventures with a threshold of AI content should mandate worker equity participation. The goal should be to spur millions of AI-enabled small and midsize firms across the real economy—not concentrate wealth narrowly in a few large firms in the technology or financial sectors.

Fourth, dynamism requires ownership. Much of the country’s wealth remains tied either to wages or to housing assets dependent on continued income flows. We cannot wait to broaden participation in the AI economy until after its gains are already concentrated among a connected few. In the AI era, data itself is becoming a productive asset class. Americans should not participate in this economy solely as consumers or sources of extraction. A modern AI Homestead should establish baseline rights that allow citizens and communities to participate in the economic value created from the data they generate. Congress should therefore establish a framework for personal data rights that treats data not merely as a privacy issue, but increasingly as an economic asset.

These elements are interlinked. You cannot own what you cannot access. Individual capability and labor mobility are therefore preconditions for ownership. Business formation broadens participation in wealth creation. And durable ownership increasingly depends on meaningful access to the infrastructure layer of the AI economy itself.

The final component of the AI Homestead, therefore, is community connectivity, and at the center of that debate lies compute itself. Five hyperscalers now control approximately 71 percent of global AI compute, up from 63 percent in early 2024. These companies are investing enormous sums in data centers that require land, power, water, and political consent—consent that is no longer guaranteed. In Memphis, the NAACP sued xAI over unpermitted gas turbines operating in a predominantly Black neighborhood; in Virginia, data center growth drove an 833 percent spike in capacity market prices, forcing regulators to create a new rate class to shield households, though it still forced up the average monthly bill by $11. As $98 billion in blocked or delayed projects in a single quarter of 2025 demonstrates, that friction creates leverage.

Communities that host these data centers should gain more than temporary construction jobs or modest tax benefits. They should gain an ownership stake in the future’s productive infrastructure. As a condition of major federal permitting and support, and consistent with recommendations made by others, hyperscalers should contribute a portion of compute capacity to local community compute pools that are accessible to local entrepreneurs, schools, and community colleges at below-market rates. The railroads received land grants and carried mail at regulated rates. Natural gas pipelines operate under open-access principles. The AI Homestead therefore should include a “compute commons.” This is not a tax on success—it is the same public interest obligation that prior generations of infrastructure builders recognized as essential to building public trust and political durability. The same should be true for AI.

When Lincoln signed the Homestead Act, he was not merely distributing land. He was broadening access to the productive frontier of the age. That architecture helped build the agricultural and industrial base that made the United States the world’s largest economy by the end of the century. The cost of failing to build an AI Homestead is not simply greater inequality. It is the erosion of what long made American capitalism distinct from rival systems: the belief that ordinary citizens could participate in the upside compounded wealth. America’s wages-to-wealth ladder was not an accident. It was built. The AI era necessitates building it again.

Navin Girishankar is president of the Economic Security and Technology Department at the Center for Strategic and International Studies (CSIS) in Washington, D.C.