AI CapEx boom faces terrible economics
The guest argued that the massive capital expenditure on artificial intelligence by technology giants is economically unsustainable relative to the actual revenues generated.
The argument
The guest highlighted that hyperscalers are projected to spend $3 trillion in CapEx over three years, while current AI revenues are estimated at only $70 billion to $100 billion. He contrasted this with Google's highly profitable 2004 IPO, noting that current AI leaders like OpenAI and X.ai are experiencing massive cash losses and low capacity utilization.
The thesis, stress-tested
✓ What validates it
- ✓Hyperscalers report a significant deceleration in AI-related CapEx
- ✓AI software revenue growth fails to match infrastructure spend in upcoming quarterly reports
▸ Risks discussed
- ▸AI monetization accelerates faster than expected
- ▸Hardware and compute costs decline rapidly due to technological breakthroughs
Hear it yourself
"all these mag companies in their history is 1 and a half trillion dollars. Think about it. Now they're talking about 3,000,000,000,000 in just three years. So these bills are created with not much capital. You're spending a trillion dollars a year, and the revenue on AI talk about maybe $7,080,000,000,000. When Google went public, I remember it was 2004. They were worth 50,000,000,000 market cap, but they're generating 700,000,000 of free cash flow. Very cash out of business. 20% plus operating margins. If you look at OpenAI and Tropic, SpaceX, a whole different league."
AFFILIATE LINK · ZORTIX MAY EARN A COMMISSION · NEVER A RECOMMENDATION TO TRADE