The AI industry needs staggering sums to survive through 2030, and the numbers don't add up. OpenAI plans to spend $750 billion on compute costs by end of 2030. Hyperscalers, including Meta, Google, Microsoft, Amazon, and Oracle, already carry $1.65 trillion in off-balance-sheet obligations and debt. This is not confidence. This is number-blindness dressed as strategy.
Ed Zitron's latest piece in 'Where's Your Ed At' runs over 7,000 words and targets a specific question: how much capital do the key AI players actually require through 2030. He covers the lead hyperscalers, the two dominant AI chip makers (NVIDIA and Broadcom), the main neoclouds (CoreWeave, Nebius, IREN), and the top AI labs (OpenAI, Anthropic, SpaceX). The analysis is built on consensus analyst estimates and original research. The comparison to Zuckerberg's $80 billion metaverse collapse is not decorative. It is the thesis.
The value of reading this in full is not the conclusion. It is the company-by-company breakdown of what each player needs versus what the market can plausibly sustain. Zitron has spent years arguing this bubble will collapse. This piece shows the mechanism, in hard numbers, before it happens.
[READ ORIGINAL →]