Michael Burry just called BS on the AI slowdown.
Michael Burry argues the AI slowdown story is wrong. The real limits are power, chips and labour, not demand.
The Big Short investor, posting as Cassandra Unchained, called OpenAI, Anthropic, and hyperscaler execs self-serving for talking about slowing things down. The money beats are the ones that stick. Competition is coming up fast, so a slowdown helps incumbents. IPOs need hype and puffery, and we are so awesome it could become dangerous is exactly that. And the safety talk can cover real slowing growth while listings slip.
He also claims LLMs are not AI and will not be AGI. That is his line. I am not buying that beat. The models are useful. Builders know it.
What I will take seriously is the cashflow read next to the physical wall. Demand is racing. Power, transformers, interconnection, and skilled trades are not. Transformer lead times stretch past 160 weeks. PJM-class queues run years. Data center labor is a wall even when capital is ready. Capex keeps landing before the last wave pays itself back.
Read the lab fine print too. Amodei said pacing does not mean halting model training. Capability cadence can slow while training continues. That dial matches Burry's IPO and growth cover points more than a full freeze.
Builders keep shipping. Just stop mistaking a capacity and cashflow story for pure altruism.
Curious which beat sticks for you: incumbents, IPO puffery, or growth cover.
Sources
- Michael Burry on X
- Business Insider: Michael Burry says AI leaders' calls for slowdown are self-serving
- Dario Amodei: We Must Pace the Frontier
- TechTarget: AI CEOs call for slowing AI boom as billions flow to data centers
- Reuters: US power companies scramble to secure equipment as data center demand surges
- Tom's Hardware: AI data center boom hits a human bottleneck
- Herald Corp: Big Tech's AI slowdown push masks IPO math and competitive calculus