You know what's interesting? Everyone still remembers how much Michael Burry made in 2008 – we're talking $100 million personal profit plus $700 million for his investors through Scion Capital. The guy literally called the subprime crisis and became a legend. But here's the thing nobody wants to admit: his track record since then has been pretty messy.

Burry's been throwing out bearish predictions left and right while markets kept climbing. He finally shut down his hedge fund last year, basically admitting he couldn't keep up with the market momentum. That's a huge deal if you think about it.

Now he's pushing this whole "AI is Dot Com 2.0" narrative, and honestly, the data just doesn't back it up. Let me break down why his thesis is falling apart:

First, he's claiming tech giants are cooking their books with depreciation schedules to pump earnings. But reality check – while GPU lifespans are shorter than traditional servers, most AI infrastructure actually lasts 15-20 years. Plus, older GPUs don't become worthless overnight. They're perfect for inference tasks, running models for end users. That's real value that Burry's overlooking.

Second, Burry's worried about cash flow getting crushed by massive CAPEX spending. Except... it's the opposite. Hyperscalers are seeing cash flow from operations surge. Alphabet alone jumped from under $100 billion to $164 billion recently. And margins? They're expanding dramatically. Companies scaling AI are reporting $3 in returns for every $1 invested. Even the latest wave – agentic AI – is helping enterprises cut costs by 25% or more. That's not a sign of a bubble.

Third, he's comparing NVIDIA to Cisco at the 2000 peak, saying valuations are insane. But when Cisco topped in March 2000, its P/E was over 200. NVIDIA's sitting at 47 right now. That's not even close to bubble territory.

What's actually happening? GPU rental prices for the H100 just spiked about 17% since mid-December, showing continued scarcity and strong demand. This is bullish for AI infrastructure plays and energy companies solving the power bottleneck.

Options traders are clearly betting on this too. We saw a $9 million whale bet on NVIDIA March calls, and Bloom Energy saw serious call buying ahead of earnings. BE even jumped 8% while the broader market was weak, breaking out of a solid weekly bull flag.

Look, Burry's contrarian reputation is solid – nobody can take that away from him. But when it comes to the current AI landscape, the evidence just stacks against his doom-and-gloom narrative. The infrastructure demand is real, efficiency gains are real, and the cash flows are real. Sometimes being a legendary contrarian means you eventually get it wrong.