So here's something that's been bugging me about the michael burry stocks narrative everyone keeps pushing. The guy literally called the 2008 crash and made a fortune doing it - that's legendary status, right? But here's where it gets messy. Since that 2008 windfall, Burry's track record has been pretty rough. He's been consistently early on his bearish calls, watching markets surge while he's sitting on the sidelines. Eventually he even shut down his hedge fund because he couldn't align with where the market was actually going.

Now he's back with a new thesis - AI is the next dot-com bubble waiting to pop. And look, I get the appeal of the contrarian narrative. But when you actually dig into the data, michael burry's current argument falls apart pretty quickly.

His core claim is that big tech companies are basically cooking the books. He points to Meta, Microsoft, and Alphabet using aggressive depreciation schedules to hide the true cost of AI infrastructure. Sounds damning until you realize that while GPUs do depreciate faster than traditional servers, the actual useful life of most AI infrastructure is 15-20 years. Plus, older GPUs aren't worthless once new chips drop - they get repurposed for inference work. That's real economic value.

Then there's the cash flow argument. Burry warns that massive CAPEX spending will drain liquidity and crush profitability. But the actual numbers tell a different story. Alphabet's operating cash flow jumped from under $100 billion to $164 billion in 2026. That's not a company in distress - that's a company generating more cash than ever. And margins? They're expanding dramatically. Companies deploying AI are reporting returns of over $3 for every $1 invested. The newest wave - agentic AI - is helping enterprises cut costs by 25% or more.

The valuation comparison is where Burry really loses me. He's comparing NVIDIA to Cisco in 2000, arguing history will repeat. But Cisco's P/E ratio topped out above 200 back then. NVIDIA's current multiple sits at 47. That's not even close to comparable. You can't just cherry-pick historical parallels and ignore the actual numbers.

What's actually happening in the market right now tells the real story. H100 GPU rental prices have spiked roughly 17% since mid-December alone. That's pure supply-demand signaling - the market is hungry for AI compute and it's not slowing down. This scarcity is bullish for GPU infrastructure plays and energy providers that solve the power bottleneck.

Options traders are clearly positioning for this too. We've seen major call buying in companies like Bloom Energy and NVIDIA, with some whale-sized bets hitting $9 million on March calls. That's not retail FOMO - that's sophisticated money betting on sustained AI infrastructure demand.

Look, michael burry stocks and his contrarian positioning used to be the trade to fade the consensus. But sometimes the consensus is just right. The AI boom has real economics behind it, real cash flow generation, and real efficiency gains. The data doesn't support a dot-com 2.0 scenario. If anything, the signal from GPU pricing, cash flow expansion, and margin improvement suggests we're still in the early innings of this cycle.