00:00The first question I have is very basic. I'm not sure if I understand.
00:03Does that mean it's a total of $35 billion in deal or are they expanding the $14 billion deal now
00:08to $21 billion?
00:10You know, I think it's expanding the deal from $14 to $21, but still it's a lot of money.
00:14It doesn't matter if, you know, either way it's a big commitment by Meta to say that, you know,
00:20we are going to invest more in either training the models and being competitive in that area.
00:26And they're going out to new clouds such as CoreV for it.
00:29But we've already seen how Nebius has done well in the last, you know, few months.
00:33And, you know, it's the same extension of that concept that if you can't build it in-house,
00:38you're going to go out to these, you know, new clouds to get the GPU as a service work that
00:43you need.
00:44Well, and they're still trying to build it in-house.
00:46Anurag, just how constrained are these hyperscalers still?
00:51How much more do they need?
00:53As Matt was pointing out, it's like every other week we get a headline of some crazy multiple tens of
00:58billions
00:58of dollars, deals being done for these data centers.
01:02See, everybody is constrained.
01:04It's not just the hyperscalers.
01:05Even Nebius and CoreV, they are also constrained in it.
01:08But you have to remember one thing.
01:10And we just learned again from the AWS, the Amazon letter today.
01:14You know, they are seeing a level of demand that they have not seen before.
01:18And this is something we go back on our work that if you look at the backlog of deals,
01:23we have seen, you know, you could say cloud commitments north of $1 trillion.
01:27I think it's north of $1.5 trillion at this point from everybody.
01:31You know, Microsoft, it's over $600 billion they have on their balance sheet that needs to drop into,
01:35you know, sales.
01:36So everybody is getting those backlog orders either from just the model companies
01:42or they're also coming from enterprises that are infusing more AI into their daily work.
Comments