Skip to playerSkip to main content
  • 22 hours ago
Transcript
00:00With respect, do investors have short memories? Because Meta told us in the prior quarter
00:05that capital expenditures would be higher in the next fiscal year. They used a different
00:09language this time around, which was notably larger in fiscal 26 and 25. But maybe it's what
00:15they didn't say. They didn't tell us what top line or bottom line growth directly from AI
00:19investments looks like. What's your read? Well, actually, for Meta, the investments look more
00:25riskier than they are for Microsoft or Amazon, who do have these data centers. And they say
00:31that if we overinvest, we will have the option to sell this excess capacity. And if you're
00:36looking at the market reaction and the mixed reaction there, we also think that macroeconomic
00:41context here is very important. Important in the sense that just before these earnings
00:46came out yesterday, the Federal Reserve said that they're not sure that they will be cutting
00:50the interest rates by 25 basis points. So I think that that also dampened the mood just
00:56before the earnings came in. Looking at the spending, these companies must spend in order
01:01to keep up with demand. And Microsoft has been very clear about the fact that they have not
01:05been able to catch up with the significantly higher demand. So they must invest. They do have
01:10the free cash to invest. If they do not invest and miss the turn, the AI turn, that would be worse
01:16for investors than seeing them over-investing. So I think that in the context of the actual graph,
01:23these companies must invest. And it doesn't really bother us as long as there are no concerns
01:28about oversupply. And this is not the case right now.
01:31Look, Meta said if they have too much compute, they can start selling it to others. So is this
01:36an air pocket? Is this just where people buy into weakness? Or is there going to be a significant
01:40pullback and questioning of the market writ large if the Fed isn't going to be cutting come December?
01:44Well, I think that the Fed situation could eventually lead to a certain pullback in valuations.
01:52But I don't think that that's going to be a bubble pop kind of a market reaction because these
01:57companies must invest now in order to make sure that they don't hit into capacity constraints one
02:03year, one and a half year from now. I believe that investors understand that. There is a report
02:08showing that or claiming that the data capacity computing needs will be doubling every nine to 18
02:15months due to our AI application. So these companies must keep up with that pace. And again,
02:21I believe that these spending plans have been almost entirely put out there. So they have been revised
02:28higher. But this is not a surprise coming in right now. Looking at the deals and the strength of the
02:34demand that we see in the sector, these companies are out there trying to keep up with the demand.
02:40And I believe that, again, as the executives are also saying, not keeping up with the demand or
02:45risking to head into capacity constraints is worse than over investing right now.
02:52What's interesting is, of course, perhaps just to show the amount of spend that was necessary,
02:57Meta comes to market with a ginormous $25 billion bond sale. And talk to us about demand
03:03across assets for AI-related companies. Shall we pull back in equities that are near record highs?
03:08How is demand looking for the bond side of the equation?
03:12Well, I think the bond side is also looking well, especially right now. We have seen that the bond
03:17sales have been quite interesting for these companies. I believe that in the actual environment
03:22as well, the bond sales are going to be interesting. Some investors who are not necessarily willing to
03:26take the equity risk at the higher valuations, they will be plugging into the bond side of the market.
03:31You are not expecting them to move separately. You expect them to move in tandem. But on the bond
03:38side, the potential of rise is less and the risk that we're taking for it is also less. So this is
03:44another way of taking exposure and positive exposure to these companies without however taking the risk
03:50of the equity fluctuations.
03:53Ipek, if I take the earnings statements and call transcripts of all the tech earnings we have so far,
03:59run them through ChatGPT and ask ChatGPT to cross-reference for a theme that is not capital expenditures,
04:06what would I get? What else do you see that they have in common between them?
04:10Well, the couple of expenditures are actually rising. I don't know. I haven't asked ChatGPT what he
04:17thinks about it. What we see is that this is growing exponentially and there is investment
04:22worries about that. But if you do not invest today, again, a year and a year and a half from now, you
04:27might hurt into bigger capacity constraints. What's happening right now is if today, for
04:32Max 1%, for example, you are not able to catch up with the demand, then tomorrow it's going to be
04:37harder because it will just be divergingly because demand is going to be exponentially
04:42rising. And that's the main issue. That's why these companies are investing so badly or so
04:47highly. And this is, I think, what ChatGPT will tell you, because as the AI applications will be up
04:53and running, demand for competing is going to rise exponentially. And this is exactly why NVIDIA is also
04:59not only concentrating today on the AI applications, but all on the networks and intelligent networks
05:05as well to make sure that even networks and the pipelines will be able to handle all that data
05:11flow down the road.
Be the first to comment
Add your comment

Recommended