What People Mean When They Say AI Is Killing Capitalism

By Futurist Thomas Frey

The phrase gets thrown around a lot right now. “AI is killing capitalism.” You hear it in tech circles, in economic policy debates, in late-night conversations between people who sense something fundamental is shifting but can’t quite name it.

They’re not wrong that something is shifting. But “killing capitalism” isn’t quite the right frame. What’s actually happening is more interesting and more complicated than that — and understanding it matters, because the transition we’re entering will be one of the most disorienting periods in economic history regardless of what we call it at the end.

Let me try to be precise about what’s actually breaking.

The Three Pillars That Are Cracking

Capitalism as we’ve practiced it for the past two centuries rests on three foundational assumptions that most people have never had to examine because they’ve always been true.

The first is that scarcity creates value. Things cost what they cost because they’re hard to make, hard to find, or hard to reproduce. The price of a good reflects the real constraint of producing it. This is so fundamental to market economics that most economic theory is essentially a sophisticated elaboration of this single idea.

The second is that labor creates income. People participate in the economy by trading their time, their skills, and their effort for compensation. Work is the mechanism through which most people access the prosperity that the economic system generates. This is not just an economic fact — it’s a moral and psychological one. The connection between contribution and reward is the social contract that holds democratic capitalism together.

The third is that information is expensive to produce. Research costs money. Expertise costs years. Creating original content, analysis, code, strategy — these things required human minds and human time, and that scarcity gave them value.

AI is attacking all three simultaneously.

Intelligence is becoming abundant. The cognitive work that used to require years of education and expensive professionals — writing, analysis, coding, legal reasoning, medical diagnosis, financial modeling — can now be performed at scale for fractions of a cent per query. Labor is becoming optional in a growing range of roles, not because the work has disappeared but because AI systems can do it faster, cheaper, and without the friction of human employment. And information has collapsed to near-zero marginal cost — one model can serve a million users simultaneously, each receiving a customized response, at a cost that rounds to nothing.

None of these trends is complete. All of them are accelerating.

AI breaks the link between work and income, drives costs toward zero, concentrates winners, and dissolves traditional firms. This isn’t evolution—it’s a structural rewrite of the economy.

What Actually Changes

When you pull on those three threads, four things start to unravel in sequence.

The first is the link between labor and income. This one gets the most attention and deserves it. For most of human history, value creation required human effort. Cognitive work, physical work, creative work — all of it was bottlenecked by the number of people doing it and the hours they could work. AI breaks that bottleneck. A company that used to need fifty people to do a specific cognitive task may soon need five — or one, with good tools, or theoretically none at all. That’s not a marginal efficiency gain. That’s a structural change in the relationship between human contribution and economic output.

The second is the collapse of marginal cost. When one AI system can produce content, designs, code, or analysis for millions of users simultaneously, the economics of information-based products change completely. Prices trend toward zero. Businesses that were built on information scarcity — the consultant who charged for knowledge, the agency that charged for execution — find their pricing power evaporating. The value migrates from the production of the output to the ownership of the system that produces it.

The third is the extreme concentration of winners. AI scales instantly and globally. The best system in any given category doesn’t compete with other systems by being slightly better — it dominates, because the marginal cost of serving another user is essentially nothing. Winner-take-most dynamics, which were already intensifying in the digital economy, accelerate dramatically. The distance between the top of the market and everyone else grows faster than any previous technology has allowed.

The fourth is the beginning of the end for the traditional firm. Companies exist, in large part, because coordinating human effort inside an organization is more efficient than coordinating it through markets. But if AI agents can replace departments — if one person with good AI tools can do what a team of twenty used to do — the organizational logic that made large companies necessary starts to dissolve. We’re beginning to see the early edges of this already: tiny teams with AI leverage doing things that would have required substantial organizations five years ago.

Three Futures, One Transition

When I think about where this leads, three scenarios seem genuinely possible — and they’re not mutually exclusive. We may move through all of them in sequence.

The first and most likely near-term outcome is what I’d call hyper-capitalism. The system doesn’t collapse — it intensifies. Smaller teams produce bigger outputs. Companies generate more value with fewer people. Wealth concentrates around whoever owns the AI systems, controls the data those systems need, and holds the distribution channels those systems reach users through. The gap between the AI-enabled and the non-enabled becomes the defining inequality of the era — wider, faster, and more durable than anything the industrial revolution produced.

The second scenario involves governments stepping in before that gap becomes socially destabilizing. Universal basic income. Taxes on AI productivity. Public ownership of certain critical AI infrastructure. Not the end of capitalism, but a floor built underneath it — capitalism on top, redistribution running beneath. Several countries are already experimenting with versions of this. Whether they can move fast enough, and at sufficient scale, is the open question.

The third scenario is further out and more speculative, but increasingly worth taking seriously. If AI continues on its current trajectory toward genuine general capability, we may enter what I’d call an access economy — a world where most production is automated, most work is optional, and the relevant question shifts from what you earn to what you can access. Not ownership of goods and services, but access to them, potentially regardless of contribution. That’s a genuinely post-labor economy, and it would require reinventing not just economic policy but the meaning-making structures that work has always provided.

Capitalism assumes people must work. AI creates value without them. That collision breaks the system’s logic—turning institutions designed for scarcity into barriers in an age of abundance.

The Fracture Point

Here is the real tension underneath all of this, and it’s worth stating plainly.

Capitalism as a system assumes that people must work to earn. That assumption is so baked into the architecture of the system — into tax structures, social insurance programs, political legitimacy, cultural identity — that the system doesn’t know how to function without it. Remove the premise that human labor is the primary source of value creation, and the entire distribution logic of capitalism breaks down. Not immediately. Not all at once. But structurally, irreversibly.

AI introduces a world where value can be created without people. That’s the fracture point. Every other consequence flows from it.

What we’re entering now is a collision between systems built for human scarcity and a technology that creates machine abundance. When that collision happens — and it’s happening now, not in some theoretical future — every system eventually becomes the problem it was designed to solve. The labor market that was designed to connect workers to opportunity becomes the system that excludes people from prosperity. The educational system designed to prepare people for work becomes the system that trains them for jobs that are disappearing. The social contract designed to distribute the gains of capitalism becomes the system that cannot keep up with the speed at which those gains are being redistributed upward.

What This Actually Means

I want to be clear about something. AI will not eliminate capitalism. Markets are extraordinarily good at allocating resources, generating innovation, and responding to human preferences. Nothing on the horizon replaces that function. What AI will do — what it is already doing — is break capitalism’s current operating model and force it to mutate into something new.

The transition period — and I’d estimate we’re looking at ten to twenty years of genuine instability — will be characterized by three simultaneous pressures: the old system deteriorating faster than most institutions can adapt, new economic arrangements emerging faster than governments can regulate them, and the human cost of that gap falling disproportionately on the people with the least capacity to absorb it.

That’s not a prediction designed to frighten people. It’s a description of what transitions of this magnitude historically look like — and a reminder that understanding the shape of the transition is the only way to navigate it with any agency.

The question worth asking right now is not whether capitalism survives. It will, in some form. The question is what kind of system we deliberately build to replace what’s breaking — and whether we do that work intentionally, with foresight and values intact, or whether we wait for the fracture to force the conversation.

History suggests we usually wait.

This time, the speed of the fracture may not give us that luxury.