Aug 12, 2026 · 53 min
AI infrastructure demand survives chips, power and politics
CoreWeave, Supermicro Lead Tech Stocks Higher on Results
The episode tests whether booming AI infrastructure spending reflects durable demand or the early signs of a costly oversupply.
- 1Strong results from CoreWeave, Nebius and Supermicro point to persistent AI infrastructure demand despite supply and power constraints.
- 2Open and closed models are likely to coexist as enterprises balance security, continuity, specialization and cost.
- 3GPU benchmarks, futures markets and depreciation data could turn compute capacity into a more transparent financial market.
Don't miss
Carmen Lee explains how GPU spot prices, forward curves and futures could reveal an AI infrastructure oversupply before earnings do.
The brief
CoreWeave, Nebius and Supermicro are posting strong results as chip shortages, power limits and data-center opposition test the durability of AI infrastructure demand.
Michelle Weaver of Morgan Stanley argues that measurable returns from AI adoption are emerging, even as compute, labor and electricity remain bottlenecks.
The episode frames open and closed models as a hybrid market: enterprises want security and continuity, while specialized applications may favor cheaper open models.
Silicon Data’s Carmen Lee describes independent GPU benchmarks and planned futures contracts as tools for hedging compute prices and spotting oversupply before it arrives.
Venture investor Joanne Chen says AI may reorganize companies around smaller, higher-agency teams, while Suno faces the harder questions of copyright, licensing and music-market saturation.
Featuring
Nvidia Corporation
OpenAIListen to the full episode and explore every guest, topic, and moment on PodLume.