Summary

Zitron argues that the generative AI industry is structurally incapable of slowing down: the data center buildout and compute commitments by OpenAI, Anthropic, NVIDIA, and Oracle require over $2-3 trillion in annual AI revenue by 2030, a figure that current demand falls catastrophically short of. He shows that OpenAI and Anthropic make up roughly 89% of AI startup revenue, yet their combined projected 2026 revenue (~$60 billion) would need to grow nearly 500% by 2029. The cruel irony is that just as AI must accelerate, the shift to token-based billing in early 2026 is causing companies like Uber, T-Mobile, and Brex to cap spending because they can't measure any ROI. He concludes that AI's costs are kept high by design to feed a circular economy, but no honest, sustainable product underpins it. He teases a forthcoming story he claims could burst the bubble.

Key Insight

The AI industry's compute commitments require impossible revenue growth, yet token-based billing is already exposing its missing ROI and slowing demand exactly when it must accelerate.

Spicy Quotes (click to share)

• 6

  AI cannot, under any circumstances, slow down.

• 9

  This is a hysterical era perpetuated by liars, cowards, imbeciles, craven boosters and the easily-fooled.

• 6

  There's barely a few billion dollars of demand outside of two companies that lose billions of dollars a year.

• 7

  This is revenue made entirely by convincing your customers that something is true and keeping them in the dark as long as humanly possible as they run up ridiculous bills.

• 7

  These men have complete contempt for their users and customers.

• 7

  The con needed a real product at some point.

• 2

  I don't do this because of a short position. I don't have one.

• 6

  I deeply hate what is being done to the computer.

Tone

combative, profane, critical