John Foley

There was Porsche and Volkswagen.

More examples in the U.S., I think.

Also, the Rio Tinto and BHP Billets and the Miners, although it didn't happen, it was on the table, I think.

It was something that Rio Tinto thought about doing to BHP.

That's a great question.

So NVIDIA, so this company, Cerebras, tried to go public a year ago, a year and a half ago, and it didn't really manage it.

And it would have had a valuation probably of about maybe $10 billion.

And now it's got a valuation of nearly $50 billion.

And it's got OpenAI as a customer, which it didn't have before.

And it's doing, by all accounts, pretty well.

And part of that is because NVIDIA is kind of a victim of its own success.

Everyone wants NVIDIA's AI chips, but it can't make enough of them.

And maybe it doesn't really want to.

Because obviously scarcity has value.

But because everyone is desperate to get these chips, they're now starting to consider alternatives.

So Google is making its own chips.

Amazon is making its own chips.

And Cerebrus is increasingly a serious alternative.

I think we should be alert to the idea that NVIDIA has competition that didn't seem to be there before.

And we should not be surprised by that because, as you say, NVIDIA has, what, 75%-ish gross margin?