WEBVTT - Nvidia Invests $2 Billion in Nebius for New Data Center Deal 0:00:02.720 --> 0:00:10.559 Bloomberg Audio Studios, podcasts, radio news. You're listening to the 0:00:10.600 --> 0:00:14.560 Bloomberg Intelligence podcast. Catch us live weekdays at ten am 0:00:14.600 --> 0:00:17.880 Eastern on Apple, Cocklay and Android Auto with the Bloomberg 0:00:17.920 --> 0:00:21.040 Business App. Listen on demand wherever you get your podcasts, 0:00:21.360 --> 0:00:23.080 or watch us live on YouTube. 0:00:23.960 --> 0:00:27.640 All right, I listen on the Bloomberg Mobile app. 0:00:27.800 --> 0:00:29.800 Good right, I have the Black app, so I'm always 0:00:29.880 --> 0:00:33.880 listening to Bloomberg Radio on that. When you do, you'll 0:00:33.880 --> 0:00:37.239 hear stories about Nvidia, undoubtedly every single day, as it's 0:00:37.280 --> 0:00:41.440 the largest company in the world. It's investing, to today's story, 0:00:41.479 --> 0:00:46.280 two billion dollars in Nebious. It's a new data center deal. 0:00:46.880 --> 0:00:52.360 Nebbius obviously is one of those neo cloud companies. And 0:00:52.440 --> 0:00:55.160 we've got Man Deep sing here from Bloomberg Intelligence. He 0:00:55.480 --> 0:00:58.880 runs our technology coverage for BI to talk a little 0:00:58.920 --> 0:01:03.320 bit about this in Nvidia deal, as well as what's 0:01:03.360 --> 0:01:05.480 going on with Uber and Zooks, though I guess that's 0:01:05.520 --> 0:01:08.119 far less important. And indeed, what do you make of 0:01:08.160 --> 0:01:11.360 the I mean, it seems relatively small, right compared to 0:01:11.520 --> 0:01:14.880 a four trillion dollar market cap company throwing a easily 0:01:14.920 --> 0:01:16.399 two billion dollars at Nebus. 0:01:16.440 --> 0:01:17.440 But what does it get them. 0:01:17.760 --> 0:01:21.960 I mean things are changing in terms of the ecosystem 0:01:22.319 --> 0:01:25.839 and the partnerships that these companies had. Remember in Vidio 0:01:25.959 --> 0:01:29.520 plan to spend almost one hundred billion dollars with open 0:01:29.560 --> 0:01:31.880 Ai at one point, so they scale that back to 0:01:32.000 --> 0:01:34.000 thirty billion in their latest funding round. 0:01:34.360 --> 0:01:34.880 Oh guess what. 0:01:35.160 --> 0:01:39.000 They are investing a lot more in neoclouds, whether it's 0:01:39.000 --> 0:01:43.720 cor Viv or Nebus. Now, basically from an Invidia standpoint, 0:01:43.800 --> 0:01:48.320 they want more fragmentation. They don't want that cloud world 0:01:48.480 --> 0:01:52.000 to be limited to three hyperscalers. They want as much 0:01:52.040 --> 0:01:55.600 fragmentation as they can for their chips. And really, in 0:01:55.640 --> 0:01:58.840 this case, in Nebus case, they want to build an 0:01:59.000 --> 0:02:02.320 end to end in Vidio stack that Nebus is hosting, 0:02:02.360 --> 0:02:05.880 and they are getting clients to use you know, Nvidia 0:02:06.120 --> 0:02:10.320 throughout whether it's training, inferencing, and really optimize the performance 0:02:10.639 --> 0:02:13.360 to the Nvidia stack because in the end, you know, 0:02:13.560 --> 0:02:17.520 five years down the line, will the supply demand equation 0:02:17.600 --> 0:02:20.600 would be very different. Right now, everyone is supply constrain, 0:02:20.919 --> 0:02:23.959 but they are thinking five ten years ahead when chips 0:02:24.000 --> 0:02:27.120 could again become a commodity. They don't want that to happen. 0:02:27.160 --> 0:02:28.920 They want this to be more fragmented. 0:02:29.560 --> 0:02:33.079 Does this deal, this two billion dollar investment, put NEBS 0:02:33.080 --> 0:02:35.400 on the map? I mean, is this you know, game 0:02:35.480 --> 0:02:37.200 changer for that neocloud company? 0:02:37.400 --> 0:02:41.959 Absolutely? Right now it is all about raising the funds, 0:02:42.000 --> 0:02:45.280 which we heard from Oracle last night. They're not going 0:02:45.320 --> 0:02:47.880 to the band market anymore. And that's why you saw 0:02:47.880 --> 0:02:52.119 a positive reaction. So the market is very sensitive about 0:02:52.639 --> 0:02:56.920 raising more debt to finance the infrastructure build out, and 0:02:56.960 --> 0:03:00.480 that's why Oracle saw such a big pushback. I'm sure 0:03:00.520 --> 0:03:02.720 it's the same for core Weave as well. 0:03:03.080 --> 0:03:04.040 And so if. 0:03:03.840 --> 0:03:07.480 Nebus is getting two billion dollars from in Vidia, and 0:03:07.560 --> 0:03:10.639 you can say it's circular financing to buy in Vidia chips, 0:03:11.000 --> 0:03:14.360 but it alleviates that need to go to the bard 0:03:14.440 --> 0:03:17.400 market to raise a two billion dollars And so from 0:03:17.440 --> 0:03:20.160 that perspective, it does solve that problem that you know, 0:03:20.240 --> 0:03:23.080 they don't have to raise money right away. 0:03:23.240 --> 0:03:27.560 I mean, they're only one of two neo cloud companies 0:03:27.600 --> 0:03:31.239 anyone's ever heard of, right, No one knows core Weave. Yeah, 0:03:31.240 --> 0:03:32.799 beyond core Weave and nebbyus, what is. 0:03:32.720 --> 0:03:37.920 There there are there are you probably know all me 0:03:38.000 --> 0:03:42.160 and Scarlett core Weave and then and that's number one, 0:03:42.280 --> 0:03:44.880 and Nebyus is the also ran right with by the way, 0:03:44.920 --> 0:03:47.680 way to the average cost of capital at Nebus twenty percent. 0:03:48.360 --> 0:03:48.880 There you go. 0:03:49.000 --> 0:03:50.760 So, yes, they don't want to have to go out 0:03:50.800 --> 0:03:51.760 and raise money. 0:03:51.760 --> 0:03:54.960 They'd rather have a chip maker give them money to 0:03:55.000 --> 0:03:56.400 buy that chip maker's chips. 0:03:56.640 --> 0:03:58.560 Yes, that's what they cost her to a financing. And 0:03:58.600 --> 0:04:00.640 that's what has some people concerned. 0:04:00.920 --> 0:04:04.720 I mean, why doesn't in Vidia just its own neo 0:04:04.760 --> 0:04:06.440 cloud company. 0:04:06.080 --> 0:04:10.160 They are, they do have a DGX cloud offering. But look, 0:04:10.240 --> 0:04:16.000 in this case, Nvidia has new architectures every twelve months. 0:04:16.360 --> 0:04:18.640 They have already announced their Ruben architecture. 0:04:18.800 --> 0:04:22.480 So what they want is these neo clouds to have 0:04:23.400 --> 0:04:27.120 those new chips first, as opposed to an Amazon or 0:04:27.200 --> 0:04:31.240 a Microsoft. And that serves them well because these new 0:04:31.320 --> 0:04:34.320 clouds will end up signing up customers which will be 0:04:34.360 --> 0:04:37.440 long term customers, and that I think it's great for 0:04:37.600 --> 0:04:39.120 Nvidia to have that fragmentation. 0:04:39.160 --> 0:04:41.480 Can I I'm gonna throw an audible here if you 0:04:41.520 --> 0:04:43.640 don't mind, because we were going to talk about uber zooks, 0:04:43.680 --> 0:04:44.800 but I mean it. 0:04:44.760 --> 0:04:46.440 Doesn't matter in the context of Invidia. 0:04:46.440 --> 0:04:49.200 I mean, they're cool looking little things, but I don't care. 0:04:50.520 --> 0:04:53.160 What I care more about is this chip battle. 0:04:53.360 --> 0:04:53.520 Right. 0:04:53.600 --> 0:04:59.520 We knew Nvidia completely and totally dominated, still does I'm guessing. 0:05:00.720 --> 0:05:03.400 Google has a chip that's like a contender. 0:05:04.000 --> 0:05:08.160 Amazon has Trainium, and I thought that Meta had just 0:05:08.320 --> 0:05:11.440 like you know, not joined the party. But now they 0:05:11.520 --> 0:05:14.320 just showed up, Right, that's breaking news this morning at 0:05:14.320 --> 0:05:15.800 ten o'clock. What's Meta deploying? 0:05:16.200 --> 0:05:19.080 And look, that's the big risk for an Nvidia is 0:05:19.160 --> 0:05:23.080 all these hyperscalers don't want to spend you know, thirty 0:05:23.120 --> 0:05:26.200 forty billion dollars a year on buying Nvidia chips, so 0:05:26.320 --> 0:05:30.919 that's where they will continuously try and develop their own chips. Now, 0:05:31.160 --> 0:05:34.760 Meta is way behind at Google TPU. Google TPU is 0:05:34.800 --> 0:05:38.320 still you know, deployed. They have seven versions of their chips. 0:05:38.560 --> 0:05:41.039 Meta is still in the initial stages of building. But 0:05:41.160 --> 0:05:44.520 with the scale that Meta has, that means down the 0:05:44.560 --> 0:05:48.040 line it will be less off a purchase for Invidia, 0:05:48.120 --> 0:05:50.480 and that's the risk that Nvidia wants to avoid. 0:05:50.520 --> 0:05:53.760 Here, Stay with us. More from Bloomberg Intelligence coming up 0:05:53.839 --> 0:05:54.280 after this. 0:05:57.960 --> 0:06:01.679 You're listening to the Bloomberg Intelligence podcast. Catch us live 0:06:01.760 --> 0:06:04.839 weekdays at ten am easterne on apple Cockplay and Android 0:06:04.839 --> 0:06:08.159 Auto with the Blueberg Business app. Listen on demand wherever 0:06:08.200 --> 0:06:11.320 you get your podcasts, or watch us live on YouTube. 0:06:11.400 --> 0:06:14.599 Through the big gainer and the stock that everyone's paying 0:06:14.600 --> 0:06:18.000 attention to right now is Oracle biggest performer, biggest gainer, 0:06:18.000 --> 0:06:19.479 I should say in the S and P five hundred 0:06:19.520 --> 0:06:22.200 best performer up more than ten percent. And in fact, 0:06:22.279 --> 0:06:24.560 this looks like the best day for Oracle if it 0:06:24.600 --> 0:06:28.240 closes at these levels, biggest rally since it announced that 0:06:28.240 --> 0:06:31.360 three hundred billion dollar open aideal. Let's bring in on 0:06:31.520 --> 0:06:33.560 rag Rana. He is our go to guy in all 0:06:33.560 --> 0:06:35.440 things tech because he is our tech analyst here a 0:06:35.440 --> 0:06:39.640 Bloomberg Intelligence and Oracle did report earnings. Were the earnings 0:06:39.680 --> 0:06:43.039 that solid to justify this massive ten percent in advance? 0:06:43.400 --> 0:06:45.000 You know, I think when you go back and see 0:06:45.080 --> 0:06:49.160 when when Microsoft and Amazon and Google reported, they all 0:06:49.200 --> 0:06:53.160 talked about increasing capics and that spooked investors out. And 0:06:53.320 --> 0:06:56.960 I mean, I guess Oracle had the foresight to say that, Listen, 0:06:57.040 --> 0:07:00.440 we added a lot more backlock to our you know, 0:07:01.000 --> 0:07:03.880 balance sheet, but we're not raising capex. I think that's 0:07:03.920 --> 0:07:07.599 the big difference right now. That's what the market is less. 0:07:08.040 --> 0:07:10.680 I think concerned about that maybe they're going to be 0:07:10.840 --> 0:07:14.040 not as aggressive as the others in terms of capital expender. 0:07:14.160 --> 0:07:16.440 So it's relief that it did not say it's full 0:07:16.480 --> 0:07:17.080 your outlook. 0:07:17.320 --> 0:07:17.920 Yeah. Yeah. 0:07:18.000 --> 0:07:20.120 And the other thing you have to say is when 0:07:20.160 --> 0:07:23.080 you look at their press release, they talked about two things. 0:07:23.360 --> 0:07:27.120 One was for the customers coming in, they're going to 0:07:27.120 --> 0:07:29.400 ask them to prepay some of that stuff or they're 0:07:29.400 --> 0:07:31.920 going to ask them to bring their own GPUs because 0:07:32.040 --> 0:07:34.240 it just shows them being a little bit more disciplined 0:07:34.240 --> 0:07:35.000 on the fin out side. 0:07:35.040 --> 0:07:37.400 Do you bundle them into the other big guys or are 0:07:37.360 --> 0:07:40.080 they a separate bolt and almost to meg seven. 0:07:40.000 --> 0:07:42.560 It was a bolt on for the cloud providers a 0:07:42.640 --> 0:07:45.720 few years ago, but in the last two years, the 0:07:45.840 --> 0:07:48.800 orders that they're getting from open Ai has really catapulting 0:07:48.840 --> 0:07:50.400 them into the biggest category. 0:07:50.440 --> 0:07:53.240 People go to Oracle, what's their distinction versus going to 0:07:53.280 --> 0:07:54.400 the forty seven others. 0:07:54.600 --> 0:07:56.800 Well, the thing is there are only three big ones, 0:07:56.840 --> 0:08:01.920 the Amazon, Microsoft and Google. Frankly, the Oracle only is 0:08:01.920 --> 0:08:03.680 another one that has the capital to come up with 0:08:03.680 --> 0:08:06.520 a cloud infrastructure at par with some of the others. 0:08:07.160 --> 0:08:09.960 However, you look at what Oracle's results also showed, which 0:08:10.000 --> 0:08:12.000 is that cash flow is going to remain negative for 0:08:12.000 --> 0:08:13.920 the next few years. And this is a company that's 0:08:13.920 --> 0:08:17.800 definitely looking at its expense line. You mentioned how some 0:08:17.800 --> 0:08:19.560 cloud customers will pay for their own ships, so they're 0:08:19.600 --> 0:08:23.400 being more discerning in that regard. But they've also are 0:08:23.480 --> 0:08:26.440 making plans, according to our reporting, to cut thousands of 0:08:26.520 --> 0:08:29.880 jobs too. How much fat can they cut right now? 0:08:30.080 --> 0:08:32.199 That's a very good question because when you look at 0:08:32.200 --> 0:08:35.080 their gross margin declined quite a bit, but their operating 0:08:35.120 --> 0:08:38.520 margin or they adjusted operating margin, was only down one percent, 0:08:38.600 --> 0:08:41.240 which is basically that they cut so much in sales 0:08:41.240 --> 0:08:44.920 and marketing and general and administrative expenses to offset some 0:08:44.960 --> 0:08:45.480 of that pressure. 0:08:45.520 --> 0:08:47.120 But you're right, they cannot do that forever. 0:08:47.679 --> 0:08:49.800 They really need to get scale in and see the 0:08:49.800 --> 0:08:52.839 benefits of these cloud contracts in order to offset that. 0:08:52.920 --> 0:08:54.680 I mean, I know you don't talk to Robert Schiffman, 0:08:54.760 --> 0:08:56.640 you're not in speaking terms, but it just put out 0:08:56.640 --> 0:09:00.160 a blistering note on Salesforce and their bond deal and 0:09:00.440 --> 0:09:04.439 rag Rana. On a given thirty forty billion dollar cash call, 0:09:04.559 --> 0:09:07.680 they make four phone calls, maybe five, They bring it 0:09:07.720 --> 0:09:10.760 in three four times over subscribe. Does this party just 0:09:10.840 --> 0:09:12.920 keep on going for the man seven. 0:09:13.200 --> 0:09:14.880 The question is for how long? I mean, I think 0:09:14.880 --> 0:09:17.240 that's the big concern. But when we talk about Oracle, 0:09:17.280 --> 0:09:20.840 for example, their bounty was oversubscribed, but now they have 0:09:20.920 --> 0:09:23.240 to go in the market and raise equity also because 0:09:23.280 --> 0:09:25.240 one of the things they have said is they don't 0:09:25.240 --> 0:09:27.320 want to get rid of their investment grade ratings. So 0:09:27.559 --> 0:09:29.320 I think this is where a lot of the relief 0:09:29.360 --> 0:09:31.640 is coming for for some of the investors, that they're 0:09:31.640 --> 0:09:34.600 going to have a much more balanced approach on raising to. 0:09:34.600 --> 0:09:37.840 Finally be adults instead of being pri Madonna's. You were 0:09:37.920 --> 0:09:41.080 up at Buffalo years ago, read about these pri Ma 0:09:41.160 --> 0:09:43.840 Donnas out on the West coast, and now they're finally 0:09:43.880 --> 0:09:45.119 growing up right. 0:09:44.920 --> 0:09:47.760 Well, that's because the investors aren't behaving differently than they 0:09:47.760 --> 0:09:50.520 were before. I mean, when last year we saw Microsoft 0:09:50.600 --> 0:09:53.520 raising capecks, everybody was liking it. But just a month 0:09:53.600 --> 0:09:55.560 a month and a half ago, when they talked about 0:09:55.640 --> 0:09:58.679 raising capecks, all the cloud providers fell down. So that's 0:09:58.679 --> 0:10:00.640 a lesson for the next one. Come in and say, 0:10:00.679 --> 0:10:01.920 I don't want to get that message. 0:10:01.960 --> 0:10:03.679 Can I do an audible always? 0:10:03.760 --> 0:10:05.440 What in God's name do you read? 0:10:05.760 --> 0:10:07.280 I mean how do you keep up on this? 0:10:07.440 --> 0:10:09.719 When I was a kid, the old man would say 0:10:09.760 --> 0:10:13.360 in a stupid read this article in MIT Technology Review 0:10:13.559 --> 0:10:14.880 because he really covered a cover. 0:10:15.200 --> 0:10:16.559 What do you read to keep up? 0:10:16.720 --> 0:10:16.960 Tom? 0:10:17.000 --> 0:10:18.240 It's very hard nowadays. 0:10:18.280 --> 0:10:21.800 I think the podcast recently where Satya is speaking, or 0:10:22.080 --> 0:10:24.319 you know, when when the Google CEO is speaking, those 0:10:24.320 --> 0:10:26.520 are the ones you really have to follow because the 0:10:26.640 --> 0:10:29.400 text speed of the innovation is just so fast that 0:10:29.640 --> 0:10:30.920 you know anything. 0:10:30.600 --> 0:10:33.000 From six months ago. I mean it's just antique. 0:10:33.320 --> 0:10:35.920 And of course you're listening to these podcasts at double 0:10:36.000 --> 0:10:39.480 or triple times the speed. I'm sure when Satya talks, 0:10:40.120 --> 0:10:46.480 when the CEO Satya in Adella of Microsoft, when Satya talks. 0:10:45.880 --> 0:10:48.080 How do I just love Mark busting her check? 0:10:49.920 --> 0:10:52.240 How much does he really reveal or is it in 0:10:52.280 --> 0:10:55.240 what he does not say that you get your your 0:10:55.480 --> 0:10:58.080 most insightful thoughts on what's next? 0:10:58.320 --> 0:11:00.880 I think I really look for him to figured out 0:11:00.920 --> 0:11:02.800 how much is he going to spend more? Because at 0:11:02.800 --> 0:11:05.040 the end of the day, he is not somebody who 0:11:05.080 --> 0:11:07.200 benefits from buying in video chips. I mean, he's the 0:11:07.200 --> 0:11:10.240 one who's funding a lot of this expansion with you know, 0:11:10.320 --> 0:11:13.080 let's say, according to our calculations, he's spending in a 0:11:13.160 --> 0:11:16.800 year fifty to sixty billion dollars just on in video GPUs. 0:11:17.040 --> 0:11:18.319 That's a very big amount. 0:11:18.600 --> 0:11:21.160 So if he's doing it for the reason, I have 0:11:21.240 --> 0:11:23.120 to see that there is some AUTOAI to that. 0:11:23.400 --> 0:11:26.439 So I'm at Palladino's down at Grand Central Station at 0:11:26.440 --> 0:11:29.920 the Bourbon Bar with Joseph. He's just like old school bartender. 0:11:29.960 --> 0:11:32.120 He's like, he's not a mad manner something like. Somebody 0:11:32.120 --> 0:11:34.400 comes up to me and they go, do they really 0:11:34.440 --> 0:11:36.600 not let man deep sing an Anna rug run in 0:11:36.640 --> 0:11:39.880 the same room? Do you guys when you're at seven 0:11:39.960 --> 0:11:42.040 thirty when lexing dinner? Are you too allowed to speak 0:11:42.120 --> 0:11:43.920 now that we sit next to each other? Absolutely? 0:11:44.040 --> 0:11:44.319 Really? 0:11:44.400 --> 0:11:46.640 Yeah, Okay, I just the guy came up to me 0:11:46.679 --> 0:11:48.440 and I said, I really don't know. I mean, there's 0:11:48.440 --> 0:11:50.920 just so much you know, Scarlet, there's just so much 0:11:51.000 --> 0:11:51.720 voltage there. 0:11:51.880 --> 0:11:53.640 There's a lot of voltage there. And we tend to 0:11:53.679 --> 0:11:56.240 have them both on but at separate times because we've 0:11:56.240 --> 0:11:59.120 got to make sure that we sprinkle their expertise throughout 0:11:59.160 --> 0:11:59.560 the hours. 0:11:59.720 --> 0:12:03.079 Yeah. Well, also there's the security issues as well. How 0:12:03.120 --> 0:12:05.400 many players are there going to be in five years? 0:12:05.880 --> 0:12:08.320 See on the hyperskate cloud providers. We know of the 0:12:08.360 --> 0:12:12.200 top three for sure, it's Amazon, Microsoft neck to neck, 0:12:12.400 --> 0:12:15.280 then Google, then Oracle. Then you have the new cloud 0:12:15.320 --> 0:12:17.319 providers that's Core, Weave and Nevius. 0:12:17.559 --> 0:12:18.280 So these are the. 0:12:18.240 --> 0:12:22.120 Five I watched most closely right now because they have 0:12:22.240 --> 0:12:25.160 the capital, They are the ones with the leading chips 0:12:25.240 --> 0:12:27.400 right now, and they are the ones everybody is going 0:12:27.400 --> 0:12:30.600 through to build their applications or run their infrints or 0:12:30.640 --> 0:12:31.479 training workloads. 0:12:31.800 --> 0:12:34.600 Anright, you were saying that Oracle has made clear that 0:12:34.640 --> 0:12:37.599 it wants and maintain its investment grade credit rating, and 0:12:37.640 --> 0:12:39.319 of course there were some concerns about that, and that's 0:12:39.320 --> 0:12:41.840 why we saw the CDs the credit to fault swaps 0:12:42.200 --> 0:12:44.640 climb in recent weeks, although it's come back down a 0:12:44.640 --> 0:12:44.960 little bit. 0:12:45.040 --> 0:12:45.280 Here. 0:12:45.679 --> 0:12:47.360 How convinced are you that they can do that? 0:12:47.800 --> 0:12:49.800 Yeah, And I talk to Rob Schifman all the time 0:12:49.840 --> 0:12:52.800 as well, so here I are on the same page 0:12:52.800 --> 0:12:55.640 that they're very careful about that investment gate rating and 0:12:55.800 --> 0:12:57.720 whenever they come up with the bond deal it gets 0:12:57.880 --> 0:12:59.000 covered up very quickly. 0:12:59.280 --> 0:13:01.760 How much does that to equity investors? 0:13:01.920 --> 0:13:03.840 I mean it does matter because you know you don't 0:13:03.840 --> 0:13:06.480 want to be financially irresponsible for a company like this 0:13:06.600 --> 0:13:09.320 that houses one of the most important software products, their 0:13:09.400 --> 0:13:12.120 database business, that is really the cash cow for them 0:13:12.160 --> 0:13:15.280 that can allow them to expand. What we saw last 0:13:15.360 --> 0:13:17.920 night was the expansion is going to be there, but 0:13:18.040 --> 0:13:21.280 maybe with a little bit more more measured means rather 0:13:21.320 --> 0:13:22.600 than going you know, all in. 0:13:23.480 --> 0:13:26.320 Stay with us. More from Bloomberg Intelligence coming up after this. 0:13:29.679 --> 0:13:33.400 You're listening to the Bloomberg Intelligence podcast. Catch us live 0:13:33.480 --> 0:13:36.520 weekdays at ten am Eastern on Apple, Coarplay and Android 0:13:36.559 --> 0:13:39.880 Auto with the Bloomberg Business app. Listen on demand wherever 0:13:39.920 --> 0:13:43.080 you get your podcasts, or watch us live on YouTube. 0:13:43.640 --> 0:13:46.240 All right, let's go to another one of our talented analysts, 0:13:46.280 --> 0:13:48.959 Herman Chan. He is our senior analyst covering US banks 0:13:49.000 --> 0:13:52.120 and Herman, there's a big headline today that got everyone's attention, 0:13:53.200 --> 0:13:56.760 JP Morgan marking down the value of some private credit loans, 0:13:57.040 --> 0:14:00.400 mainly to software companies, in the latest sign of stress 0:14:00.520 --> 0:14:03.199 in private credit. I guess it's not a surprise if 0:14:03.200 --> 0:14:05.680 you recall that Jamie Diamond, the CEO, had talked about 0:14:05.720 --> 0:14:07.640 the possibility of more cockroaches. 0:14:07.720 --> 0:14:10.599 That's right, So we view this as a prude and 0:14:10.760 --> 0:14:15.240 risk management measure by JP Morgan. Basically, by reducing the 0:14:15.360 --> 0:14:18.000 value of these loans, it means like the private credit 0:14:18.000 --> 0:14:21.360 companies can borrow less from JP Morgan, so it reduces 0:14:21.400 --> 0:14:25.440 their exposure in the case that there's some more volatility 0:14:25.480 --> 0:14:27.480 ahead for the private credit folks. 0:14:29.280 --> 0:14:33.160 So to me, what was the most interesting about this 0:14:33.280 --> 0:14:37.120 news was not that JP Morgan marked some of these 0:14:37.520 --> 0:14:40.760 credits down or some of these assets down, rather but 0:14:40.920 --> 0:14:44.760 that its competitors aren't allowed to do that. I hadn't 0:14:44.800 --> 0:14:47.040 realized before I read the FT piece and then the 0:14:47.040 --> 0:14:55.160 Bloomberg cover that that JP More that JP Morgan's competitors 0:14:55.200 --> 0:14:58.600 and probably a lot of other lenders that lend to 0:14:58.680 --> 0:15:01.680 smaller direct lenders have it in their covenants that they're 0:15:01.720 --> 0:15:06.600 only allowed to mark their assets during like change credit 0:15:06.640 --> 0:15:07.120 event or. 0:15:07.200 --> 0:15:10.800 Exactly like open enrollment. Right, that's insane. 0:15:11.440 --> 0:15:13.040 Is that a reason that we see so many of 0:15:13.080 --> 0:15:17.000 these private credit assets going from like ninety seven to zero. 0:15:18.760 --> 0:15:21.480 It just goes to show that JP Morgan is the 0:15:21.520 --> 0:15:24.800 big kahoona and they can dictate credit terms, whereas others 0:15:24.880 --> 0:15:29.240 may not have the same capacity. So having those advantageous 0:15:29.280 --> 0:15:32.880 credit terms helps them protect themselves in the event that 0:15:32.960 --> 0:15:38.440 we have potential loss We've seen headlines of and expectations 0:15:38.480 --> 0:15:42.280 and predictions of fifteen percent potential losses in the credit 0:15:42.640 --> 0:15:48.800 software books for these private credit companies. So I think 0:15:48.800 --> 0:15:50.640 banks will continue to try to get ahead of it. 0:15:51.600 --> 0:15:53.840 And then you speak to some of the credit terms 0:15:53.840 --> 0:15:56.560 and conservative underwriting at JP Morgan. I think that's one 0:15:56.560 --> 0:15:59.520 of the reasons why Western Alliance was dinged a little 0:15:59.560 --> 0:16:02.240 bit with the Jeffries news because they thought they had 0:16:02.360 --> 0:16:05.680 pristine credit terms, but then Jeffries eventually just backed out 0:16:05.680 --> 0:16:06.520 of paying back the loan. 0:16:07.080 --> 0:16:09.240 So the backstory here, if we take a step back, 0:16:09.280 --> 0:16:11.280 is that these Wall Street banks like a JP Morgan 0:16:11.400 --> 0:16:14.000 are not making a lot of private loans directly because 0:16:14.000 --> 0:16:17.520 they de risked following the Great Financial Crisis. Regulators were 0:16:17.560 --> 0:16:21.240 breathing down their necks. But they are exposed indirectly because 0:16:21.320 --> 0:16:24.720 they lend to private credit funds. Do we have a sense, 0:16:24.760 --> 0:16:27.960 herman or is this still kind of something vague of 0:16:28.000 --> 0:16:30.600 just how exposed Wall Street banks are to private credit funds. 0:16:30.720 --> 0:16:34.680 Yeah, we have some data that the Federal Reserve puts 0:16:34.680 --> 0:16:38.720 out every quarter and we calculate it. So it's about 0:16:39.080 --> 0:16:41.200 for the banks that I cover, it's about fifteen to 0:16:41.240 --> 0:16:44.360 twenty percent of their total loan portfolio is to non 0:16:44.440 --> 0:16:48.200 bank financial institutions, which private credit is a subset of. 0:16:48.360 --> 0:16:54.280 So it's a growing piece of the banking industries pie 0:16:54.800 --> 0:16:59.280 and it's really last year was the sole growth driver 0:16:59.520 --> 0:17:02.320 for banks. For JP Morgan as an example, they grew 0:17:02.360 --> 0:17:05.640 their non bank financials books seventy eight percent year every year. 0:17:05.720 --> 0:17:11.240 So it's it's a hefty position for for banks across 0:17:12.440 --> 0:17:16.080 large and regional banks. And really that's because that's where 0:17:16.119 --> 0:17:19.080 the growth was because of this financial arbitrage where banks 0:17:19.119 --> 0:17:24.560 are have lower risk weighted assets on these loans, so 0:17:24.880 --> 0:17:27.720 it helps from a capital treatment standpoint to lend to 0:17:27.800 --> 0:17:29.280 these non bank financials. 0:17:29.560 --> 0:17:34.960 So you say that JP Morgan's book there grew what 0:17:35.160 --> 0:17:38.360 seven or eight per seventy eight seventy eight percent, that's 0:17:38.359 --> 0:17:41.600 a massive growth. Are there other banks that also grew 0:17:41.640 --> 0:17:45.719 those loan portfolios massively and maybe are more reluctant to 0:17:45.720 --> 0:17:46.359 mark them down. 0:17:46.520 --> 0:17:50.880 It's across the board, right, so Wells Fargo was particularly 0:17:50.920 --> 0:17:54.520 aggressive in the fourth quarter. The banks that have larger 0:17:54.800 --> 0:17:58.680 markets businesses like a JP, like a b of A, 0:17:58.960 --> 0:18:01.520 like a city have grown much faster than some of 0:18:01.520 --> 0:18:04.560 the regionals, but the regionals have also participated as well. 0:18:04.600 --> 0:18:07.280 Yeah, I think about the headline just last month about 0:18:07.320 --> 0:18:09.840 Bank of America committing twenty five billion dollars of its 0:18:09.880 --> 0:18:12.600 own cash to private credit deals. In that instance, it's 0:18:12.600 --> 0:18:15.440 making those loans directly as opposed to indirectly through private 0:18:15.440 --> 0:18:16.040 credit funds. 0:18:16.160 --> 0:18:16.920 Right, That's right. 0:18:17.000 --> 0:18:21.560 So that's there is a bit of a regulatory arbitrage, 0:18:21.560 --> 0:18:24.960 as I mentioned before, where you have to hold less 0:18:25.000 --> 0:18:27.719 capital when you lend to private equity or at private 0:18:27.720 --> 0:18:29.160 capital versus doing the loan. 0:18:29.680 --> 0:18:32.280 Does you feel like, I don't know, two thousand and seven, 0:18:32.359 --> 0:18:34.760 when different firms were trying to get in on subprime 0:18:34.800 --> 0:18:38.040 mortgage after the boom of the growth had already been seen, 0:18:38.080 --> 0:18:40.000 and there's a little bit of FOMO driving. 0:18:39.800 --> 0:18:44.280 Activity leading the witness, it does feel like, not in. 0:18:44.240 --> 0:18:47.040 A quarterable we are feeling a bit of froth in 0:18:47.080 --> 0:18:51.600 the markets. We're we're seeing some stress, but. 0:18:51.560 --> 0:18:52.040 We have it. 0:18:52.280 --> 0:18:55.320 We're not at that level where we think things are 0:18:55.359 --> 0:18:59.160 seizing up at this point, and private credit as a whole, 0:18:59.520 --> 0:19:03.760 the the industry is about one point seven trillion dollars, 0:19:03.840 --> 0:19:08.200 so it's it could be fairly absorbed within the industry. 0:19:08.680 --> 0:19:11.040 You know, you have a bank's like JP Morgan, BFA, 0:19:11.080 --> 0:19:14.639 Wells Fargo City themselves, their entire balance sheets over a 0:19:14.640 --> 0:19:15.359 trillion dollars, So. 0:19:15.600 --> 0:19:17.399 Not two thousand and seven, but maybe, like I don't know, 0:19:17.440 --> 0:19:18.960 two thousand and six. 0:19:19.000 --> 0:19:20.720 Well, I mean, the question is what's the fallout? 0:19:20.800 --> 0:19:23.640 Right, I know it's only one point seven trillion, and 0:19:23.880 --> 0:19:28.640 frankly software is probably less than thirty percent of that, 0:19:28.760 --> 0:19:33.159 so you're talking about five hundred billion, max. But the 0:19:33.240 --> 0:19:36.080 question is what kind of fallout do you have because 0:19:36.760 --> 0:19:38.720 a lot of these banks don't have the kind of 0:19:38.760 --> 0:19:42.560 direct lending approach that Scarlet's talking about BFA. Rather, they'd 0:19:42.640 --> 0:19:47.280 lend money to the BDC and then the BDC rented out. 0:19:47.320 --> 0:19:49.520 And then those BDCs, when they're looking. 0:19:49.240 --> 0:19:53.040 At everybody headed for the gates, sell off their best 0:19:53.080 --> 0:19:55.080 assets first so that they can say we got ninety 0:19:55.080 --> 0:19:57.280 seven cents on the dollar, and then they're left with 0:19:57.359 --> 0:19:58.400