Zaid

And it looks like Nvidia has already committed pretty much all their free cash flow on this capacity.

The point that Burry is trying to make is Nvidia is being forced to place these massive orders well before they actually know what demand will look like.

Now TSMC is requiring these longer term contracts because Nvidia chips have gotten so complex that TSMC has to build custom fabrication capacity just for Nvidia.

So essentially Nvidia is placing huge bets on future demand without knowing if that demand will be there.

Burry compares this to what happened with Cisco back in the dot-com bubble.

Cisco was the picks and shovels play of the late 90s internet boom, just like Nvidia is for AI today.

At the time, Cisco extended massive purchase commitments, expecting 50% annual growth to continue.

Well, when enterprise IT spending collapsed, Cisco had to write down about 40% of its supply chain obligations and Cisco's stock tank.

So I think Burry thinks that could also happen to Nvidia.

you know, if there's a pullback on AI spending that some think will happen, maybe NVIDIA could suffer the same fate that Cisco did back during the dot-com crash.

And that's the key question here, right?

Will there actually be a pullback in AI spending, especially from the hyperscalers?

As I mentioned earlier, 50% of NVIDIA's revenues come from Meta, Microsoft, Google, and Amazon.

And these companies are expected to spend close to $700 billion combined on AI infrastructure this year.

But all this CapEx spending is actually wrecking havoc on the balance sheets of

these companies.

Google, Microsoft, Amazon, and Meta are all expected to see a noticeable decline in free cashflow this year.

In fact, Amazon is expected to go into negative free cashflow from all their CapEx spending.

So it's possible the shareholders of these companies might start pushing back and asking whether this CapEx spending is actually worth it.

In fact, the stock price for all the hyperscalers is in the red this year.