AI is slowing down
Points and comments are a snapshot, not live.
AI infrastructure buildout requires $2+ trillion annual revenue by 2030, but current demand cannot justify the costs being incurred.
Ed Zitron argues that the AI industry faces an unsustainable financial structure. He cites 190GW of planned data centers costing $9.5 trillion to $15 trillion, with banks already concerned about issuing $500 billion to $1 trillion annually in data center debt. OpenAI projects $852 billion in burn through 2030 with $770 billion in compute commitments; Anthropic has $375 billion in commitments and must reach $174 billion annual revenue by 2029. Combined, the two companies generate roughly 70-90% of all AI compute demand but need to double revenues every year through 2030 while raising hundreds of billions in additional funding. Meanwhile, actual demand from enterprises and other AI companies remains negligible: no major purchaser exists outside OpenAI, Anthropic, and hyperscalers. Zitron concludes the infrastructure can only justify itself if AI services generate over $2 trillion in annual revenue by 2030—a target that appears mathematically impossible given current adoption rates.
What commenters are saying
Commenters split on tone versus substance. The top criticism targets Zitron's abrasive, hyperbolic writing style as off-putting, even when arguments appear sound; some note he has made pessimistic predictions before with mixed accuracy. Others defend the tone as necessary pushback against equally hyperbolic AI boosterism. On substance, several commenter camps emerge: skeptics on productivity gains cite a two-year study of 4,000 teams showing no throughput increase despite claimed efficiency improvements; engineers counter that current models show real utility (though qualified as recent observations); and some note that productivity gains, even if real, may not translate to economic value creation. One commenter flags that if hyperscalers are doing equity raises instead of debt financing, as Zitron notes, that signals debt market stress. The thread's center of gravity emphasizes the gap between claimed productivity and actual business metrics.