AI infrastructure build-out is not a bubble
The current AI infrastructure build-out is fundamentally different from the year 2000 telecom bubble because there are no 'dark GPUs' and return on invested capital remains highly positive.
The argument
The guest argued that unlike the 97% of fiber that sat unlit in 2000, current GPU capacity is fully utilized. Furthermore, the capital expenditures are backed by the world's most cash-rich companies, and their return on invested capital (ROIC) has actually increased by roughly 10% since ramping up AI spend.
The thesis, stress-tested
✓ What validates it
- ✓Sustained or rising ROIC metrics in upcoming quarterly earnings for major hyperscalers
- ✓Full utilization rates of newly deployed Blackwell GPU clusters
▸ Risks discussed
- ▸Potential mismatch of near-term build-out capacity versus immediate demand
- ▸Uncertainty around whether positive ROI trends will persist with massive Blackwell spending
Hear it yourself
"And Google released a stat recently that they have seen a 150 X increase in the amount of tokens processed in the last seventeen months. So, on the one hand you've got this crazy scary sounding build out, on the other hand you actually have a bunch of usage that's happening. So are we in an AI bubble? I do not believe we're in an AI bubble today. I had, depending on how you look at it, the privilege and the misfortune of being a tech investor during the year 2000 bubble, which was really a telecom bubble. And I think it's really helpful to compare and contrast today to the year 2000. First, I think Cisco peaked at a 150 or a 180 times trailing earnings, NVIDIA's at more like 40 times."
AFFILIATE LINK · ZORTIX MAY EARN A COMMISSION · NEVER A RECOMMENDATION TO TRADE