AI compute costs outpace incremental revenue growth
The bear case presented suggests that the economic model for AI is under pressure because the cost of compute is growing exponentially faster than the productivity gains or incremental revenues it generates.
The argument
The discussion highlighted anecdotal evidence from tech investors showing that corporate compute and token spend is doubling every 45 days, while corresponding productivity gains are only rising by roughly 5%. This imbalance is expected to force corporate America to rationalize and dramatically scale back their AI compute spending, threatening the revenue growth of semiconductor and infrastructure providers.
The thesis, stress-tested
✓ What validates it
- ✓Major enterprises publicly announcing reductions or rationalizations in their AI compute budgets
- ✓A slowdown in revenue growth for semiconductor companies supplying AI chips
▸ Risks discussed
- ▸Hyperscalers successfully monetize AI services to enterprise clients
- ▸Compute costs decline rapidly due to technological efficiency gains
- ▸Productivity gains accelerate as software integration matures
Hear it yourself
"But to your point, Lance, if if costs are right now far outstripping the incremental revenue that we're getting from AI, that's gonna challenge what you were talking about earlier."
00:00 / 00:12
AFFILIATE LINK · ZORTIX MAY EARN A COMMISSION · NEVER A RECOMMENDATION TO TRADE