AI CapEx cycle remains structurally robust
The bull case argued for AI infrastructure is that Oracle and OpenAI's scaled-back data center plans do not signal the end of the CapEx cycle, as hyperscalers like Meta are eager to absorb any surplus capacity.
The argument
The guest argued that demand for compute remains structurally insatiable. He explained that the transition to persistent, 24/7 parallel AI agents running continuously will require orders of magnitude more compute than current chat-based queries, keeping the long-term demand for data centers exceptionally high.
The thesis, stress-tested
✓ What validates it
- ✓Meta or other hyperscalers officially taking over the Stargate or Texas data center capacity
- ✓Continued sequential increases in AI-related CapEx guidance from major cloud providers
▸ Risks discussed
- ▸The high cost of continuous compute may limit near-term consumer adoption
- ▸Oracle and smaller players may face cash flow constraints trying to keep up with the capital requirements
Hear it yourself
"Oracle and OpenAI end plans to expand flagship data center. Meta moves in to absorb the surplus AI data center capacity. Then Crowdstrike, they beat expectations but trade down."
00:00 / 00:14
AFFILIATE LINK · ZORTIX MAY EARN A COMMISSION · NEVER A RECOMMENDATION TO TRADE