Um, but maybe just, uh, tell me a little bit about, uh, why Nvidia, uh, started getting involved in, uh, in this, this part of the business, uh, what role you see yourselves playing there and, uh, and how, how it all is, uh, is gonna work.

Yeah.

No, that, that...

And that's a great question, and I think it ties back to the point I raised earlier, Rich, around Nvidia viewing itself as an AI infrastructure company, not just a chip company.

So I, I would, first I would, you know, kind of take, take a step back and say when Nvidia is building out its chips, we have to look at our entire downstream supply chain, right? We need to make sure that all the wafers are gonna be there, all the memory, um, elements are gonna be there in order for us to build out our end product.

There's a similar dependency that we need to be thoughtful about as well upstream.

In other words, making sure that the data centers are prepared and ready to receive the products once they land.

What's the cooling profile look like for those, and how are people adjusting to, uh, the, that faster cadence when you have to, uh, adapt your, your cooling infrastructure?

So is that new or have you been doing that for years?

NVIDIA has been a pioneer in delivering what we call accelerated computing.

The unique thing about accelerated computing is it requires very specific engagement and integration with the application itself.

And so all of those codecs libraries, those domain-specific libraries, is just another instantiation of that.

Because basically, in order to leverage computing and accelerated computing, you really need to be able to have the application, talk to that underlying hardware.

And so we've basically been doing it domain by domain.

Like if you look at some of the early use cases, they were all scientific computing bases, whether it was physics, molecular dynamics, climate and weather.

But I'm talking about just weather.