AI infrastructure spend commoditizes tech giants
The guest argued that tech companies are losing their high-margin, asset-light appeal by investing heavily in physical AI compute infrastructure, which behaves like a cyclical commodity business.
The argument
The guest compared AI data centers to oil refineries, noting they require massive physical inputs (steel, electrons, commodities) to produce 'AI compute' as a commodity. Consequently, the guest argued these companies should trade closer to cyclical multiples (10-12x) rather than their current 30-40x+ multiples.
The thesis, stress-tested
✓ What validates it
- ✓Tech multiples compress toward historical cyclical averages
- ✓Earnings reports show declining return on capital (ROC) for major cloud providers
▸ Risks discussed
- ▸AI demand could scale infinitely to offset massive CapEx
- ▸Tech companies maintain high margins through software monetization
Hear it yourself
"That means somewhere between $2.70 300 starts to have a problem. That means 40 to 50,000,000 barrels from now. And we're we're averaging somewhere between seven and ten a week. So, you know, you do the math on that. That puts you somewhere in that, you know, mid July range, which is where, you know, originally. So the and then the commercial inventories are already near critical levels. And so you're you're it won't you know, it the question is when does the market really start to price it in? Peaks the reason why I chose, July 4 is that's really when that's your peak driving season."
AFFILIATE LINK · ZORTIX MAY EARN A COMMISSION · NEVER A RECOMMENDATION TO TRADE