Skip to content

The anarchic AI race boom bubble and bust

Marxist Analysis · 2026-05-13

Core Argument

The article argues that the current boom in generative artificial intelligence is not merely a speculative bubble but a structural crisis-in-waiting, rooted in the fundamental contradictions of capitalism. It claims that the AI frenzy combines two distinct but interconnected phenomena: pure stock speculation in Big Tech shares, and a massive overaccumulation of fixed capital in data centres and infrastructure that cannot possibly reproduce its value. The central thesis is that this bubble, when it bursts, will be far more destructive than the dot-com crash because it has drawn in the entire financial system through debt, shadow banking, and circular investment schemes, and because the US economy has become dependent on AI capital expenditure to avoid recession. The article concludes that the bursting of the AI bubble could act as the trigger for a generalised economic crisis and, potentially, revolutionary upheaval — drawing a direct parallel with the 1840s railway mania and the 1848 revolutions.

Theoretical Grounding

The analysis is grounded in Marx's theory of value and his treatment of fixed capital depreciation. The article deploys Marx's concept of moral depreciation — the loss of value that occurs when machinery becomes obsolete before it has transferred its value to commodities — as the key theoretical lens for understanding why AI infrastructure investment is inherently wasteful under capitalism. This is drawn from Volume I and Volume III of Capital, where Marx discusses how newly introduced machinery "continually becomes antiquated before it has time to reproduce its own value."

The article also draws on Marx and Engels's analysis of the 1840s railway mania, particularly their argument that speculation is not an accidental feature of capitalism but a symptom of overproduction. The key theoretical move is to treat the AI bubble as a manifestation of the deeper contradiction between the productive forces and the capitalist market: capital is abundant, but profitable investment opportunities are scarce, so it pours into a single sector in a desperate, herd-like rush. This situates the analysis within the Marxist tradition's understanding of crisis as endemic to capitalism, not as an external shock.

The distinction between fictitious capital (stocks, bonds, derivatives) and real value (created by socially necessary labour in production) runs throughout the argument, though the article does not use the term "fictitious capital" explicitly. The critique of the "free market" as anarchic and wasteful, and the conclusion that only socialist planning can rationally develop the productive forces, places the analysis firmly in the classical Marxist tradition.

Conjunctural Relevance

The article is written in November 2025 and is acutely tied to the current moment. It cites specific data points that ground the argument in real economic developments:

  • Nvidia has become the first ever $5 trillion company, and the Magnificent Seven now account for over one-third of the S&P 500 — treble their share a decade ago.
  • The price-to-earnings ratio for the S&P 500 sits at 40, just shy of the dot-com bubble peak; for one-third of the index, it exceeds 50.
  • The total value of US publicly traded stocks has reached 225 percent of GDP, compared to a historical average of 85 percent.
  • AI-related capital expenditure by Google, Amazon, Microsoft, and Meta alone is projected at $750 billion over two years, with industry-wide projections reaching $2.9 trillion by 2028.
  • A MIT survey found that 95 percent of AI-driven projects fail to make it beyond the pilot stage.
  • OpenAI is projected to take in only $13 billion in revenue in 2025, while its investors are funding an estimated $1.4 trillion in AI infrastructure.
  • Margin debt stands at over $1.1 trillion, up 39 percent since April 2025.
  • AI capital expenditure is contributing more to US GDP growth than consumer spending, and without it the US economy would be in or near recession.

The article also names specific actors: Sam Altman (OpenAI), Mark Zuckerberg (Meta), Sundar Pichai (Alphabet), Jim Chanos (short-seller), Gita Gopinath (former IMF chief economist), and George Saravelos (Deutsche Bank). It references Trump's Stargate project, Trump's tariffs, and the broader geopolitical context of imperialist competition over AGI.

Where the Argument Continues

The article is itself a continuation of a running analysis on In Defence of Marxism. It references "as we have reported on recently elsewhere" regarding warnings about the AI bubble, and the broader economic context of protectionism, debt, rearmament, and crypto speculation has been treated in other IDOM articles. The argument about the 1848 revolutions as a parallel for revolutionary possibilities today is gestured at but not developed — this connects to the broader Marxist analysis of the current epoch as one of "revolution" and "organic crisis" that runs through IDOM's output and Against the Current episodes.

The article leaves underdeveloped the question of how the working class should concretely organise around the AI question — it ends with a call for revolutionary leadership and socialist planning but does not discuss specific demands or transitional programmes. This is likely treated in other IDOM articles on technology and labour, or in the RCI's programme more broadly.

Connections

  • Marx, Capital Volume I, Chapter 15 — on moral depreciation of machinery.
  • Marx and Engels, Review: May-October 1850 (Neue Rheinische Zeitung) — their analysis of the railway mania and its connection to overproduction and revolution.
  • William Quinn and John D. Turner, Boom and Bust: A Global History of Financial Bubbles — the source for the railway mania comparison.
  • Alasdair Nairn, Engines that Move Markets — on the negative aggregate returns of railway investment.
  • Ed Zitron — tech commentator and AI sceptic cited for the "vibes not returns" formulation.
  • Gita Gopinath — cited for estimates of wealth destruction from a potential crash.
  • Other IDOM articles on crypto, NFTs, debt, rearmament, and the general crisis of capitalism.

Key Quotes

  1. "When it comes to the AI world, however, as Zitron summarises: 'this is a bubble driven by vibes not returns.'"

  2. "Marx referred to this as the moral depreciation of constant capital, in contrast to the physical depreciation represented by wear and tear. 'In addition to the material wear and tear,' Marx writes in Capital, 'a machine also undergoes, what we may call a moral depreciation. It loses exchange-value, either by machines of the same sort being produced cheaper than it, or by better machines entering into competition with it.'"

  3. "What appears to the superficial observer to be the cause of the crisis is not overproduction but excess speculation, but this is itself only a symptom of overproduction." — Marx and Engels, Review: May-October 1850

  4. "In aggregate, over a very long period of time, there is no question that, for all their economic impact, the railways provided negative returns, whether you measure that in real, relative, or absolute terms. This illustrates the general truth that in the aftermath of any period of speculative excess, when companies are funded on the expectation of instantaneous stock market returns, huge amounts of capital are wasted on non-economic projects." — Alasdair Nairn, Engines that Move Markets

  5. "A market correction of the same magnitude as the dot-com crash could wipe out over $20trn in wealth for American households, equivalent to roughly 70% of American GDP in 2024." — Gita Gopinath

  6. "The only thing the capitalist market is 'efficient' at is lining the pockets of the billionaires, bosses, and banks. By all other measures, it is anarchic, destructive, and hugely wasteful."