As capital pours billions into artificial intelligence to displace human labor, the relevance of Marx’s value theory has never been greater. Marx demonstrated that surplus value derives exclusively from the exploitation of wage labor, meaning the full automation of production must ultimately destroy the basis of capitalist profit. To understand what this means for capitalism, we need to be clear about what Marx actually said about machines and value.

Ian Wright’s 2021 Cosmonaut article, Why Machines Don’t Create Value, presents itself as a defense of Marx’s value theory. In fact, it passes off Wright’s own “causal powers thesis,” the idea that labor creates value because humans are creative and adaptable in ways machines are not, under the heading of Marx’s theory. This is a retreat into philosophical idealism. Marx explicitly rejected the idea that labor has a “supernatural creative power” and stressed that labor is not the sole source of material wealth—nature and machinery are equally vital to producing use-values. Value, on the other hand, is not an innate property of human creativity; it is a historically specific social relation dictated by commodity exchange and the exploitation of wage labor.

Wright confuses a change in profit with its absolute level. He suggests humans create value because only they can innovate to get more outputs for fewer inputs. But for Marx, absolute surplus value exists whenever the working day exceeds the time needed to produce the equivalent of the worker’s wage. The capitalist buys a worker’s capacity to work at their cost of reproduction, but receives a full working day in return. Machines have no such gap; a robot’s price reflects its full productive life, and it merely yields up that value as it wears out. These are facts about production and exchange under capitalism, not about the “causal powers” inherent in machines or humans. By claiming that value is created by human “agency,” Wright reduces the critique of political economy to a transhistorical assertion that would apply equally to a serf or a slave.

This is the capitalist’s own view of machinery. Because machines raise output, the capitalist takes them to be a source of profit, confusing more goods with more value, and imagines labor can be replaced without consequence. Wright himself calls this an ideological inversion. Yet his theory concedes it in principle: he disputes only whether machines are clever enough yet.

Wright points to an excerpt from the Grundrisse to support his claims. Yet if we look at the content of that text, the argument points in the opposite direction. Marx writes that “as soon as labour in the direct form has ceased to be the great well-spring of wealth, labour time ceases and must cease to be its measure, and hence exchange value [must cease to be the measure] of use value.” Marx concludes this thought by declaring that at this point, “production based on exchange value breaks down.”

Nicholas D. Villarreal builds on Wright’s muddled foundation, most recently in his piece Marxism in the Age of AI. Villarreal’s article claims to be about AI’s “implications for the future of class struggle on Earth” as workers face total automation, but there is little in the piece on the material impact of AI on workers, or the political implications of this technology for the working class. Here, I will leave aside the interminable detours into “19th-century semiotics,” the “material negativity” of the “signifier,” and the “Landauer heat” emitted by the brain, to deal only with the brief portion of the essay that touches on political economy.

To refute the premise that full automation eliminates profit, Villarreal points out that machines cannot produce goods with a price of absolute zero because production will always require physical inputs. To be precise: in Marxist terms, the machines and the raw materials represent constant capital. Their value is merely transferred to the final product; it does not expand. What falls to zero is not the absolute price of the product, but the new value added. Because there is no human wage labor, no new surplus value is created. Under competitive pressure, the product’s price is driven down exactly to its cost of production—the cost of those transferred physical inputs—leaving zero room for profit.

Villarreal dismisses the theoretical premise of perfect competition, treating it as a flaw in modeling automation. But this assumption directly follows Marx’s method: in Capital, Marx assumes commodities sell at their values precisely so that profit cannot be explained away by cheating or monopoly. If surplus value exists even when every exchange is mathematically equal, it must come from the gap between what labor-power costs and what it produces. The same applies to machines. If a machine yields no profit when its output is sold at its value, any profit made with it under monopoly conditions must come from elsewhere. Dropping the assumption of perfect competition changes nothing about the source of value. A monopolist’s extra profit is simply taken from other capitalists.

To bypass the problem of automation eliminating surplus value, Villarreal turns to the Kalecki profit equation. He claims that surplus value “is merely the sum of capitalist consumption and investment,” arguing that profit depends purely on a difference in “nominal budgets” and “a pure accounting relationship.” By this logic, capitalists do not need to exploit labor to accumulate capital—they can generate capitalist growth indefinitely simply by spending money among themselves.

This misses the fundamental driving force of the capitalist system. Capitalists do not invest money merely to shuffle nominal budgets, nor do they invest to increase the physical mass of goods in the world. They invest for one reason: to appropriate surplus value. Villarreal conflates a physical surplus with the specific social form of surplus value. A stockpile of robot-produced physical goods is a surplus product (use-values). Every class society produces a surplus product. Surplus value, however, is the specific historical form that surplus takes under capitalism: unpaid labor time turned into money profit.

If capitalists attempt to realize profit simply by selling automated goods to each other on the market, they are playing a zero-sum game. Circulation cannot create value. Without living labor injecting new value into the system, the total pool of capital cannot expand. If robots use other robots to churn out luxury goods, the physical mass of use-values grows, but the total value in the system does not. As the value of these automated commodities falls toward the cost of their physical inputs, the systemic rate of profit vanishes.

Without the realization of surplus value, the capitalist incentive to invest evaporates, and capitalist accumulation stops. Villarreal seems to imagine that capitalists will simply ignore this, bypassing the market and continuing to invest purely to grow the physical stock of goods for their own consumption. But an economy driven by the direct production and allocation of use-values for a ruling class is not capitalism. For capitalists to smoothly transition into a class of techno-lords producing directly for their own needs, the capitalist mode of production—which is driven by the self-expansion of exchange value—would have already had to break down.

Wright and Villarreal can persist in their confused “theory” if they wish, but they should stop pretending that it has anything to do with Marxism. The implications of their ideas are reactionary. If value is simply a product of “causal powers,” and robots eventually acquire those powers, then capitalism can replace workers with machines indefinitely, even replace the whole working class. The whole basis of Marxism’s revolutionary perspective—that capitalism must inevitably break down due to its own internal contradictions—is discarded.

Villarreal comes to exactly this conclusion. Oddly, he describes this endpoint as a slave society built on universal robot labor, but at the same time treats it as a transitory depression that capitalism might survive. He predicts that the ruling class will simply manage the obsolescence of human labor via “AI-assisted methods of social control” and a “universal basic income,” and speaks of the collapse of working-class demand as a mere “transitory depression.” After this, he claims, the economy will smoothly stabilize around ruling-class consumption. But the bourgeoisie are not invincible managers who can seamlessly dial down social conflict and transition out of capitalism at will. If the exploitation of wage labor—the sole source of surplus value—were fully automated away, the capitalist system would not stabilize, but collapse. Stripping this structural inevitability from Marx’s theory leaves a working class that is obsolete, rather than one holding the unique historical power to end class society.