WEBVTT - Apollo's Torsten Slok Talks Big Tech Risk

0:00:02.520 --> 0:00:07.040
<v Speaker 1>Bloomberg Audio Studios, podcasts, radio news.

0:00:07.920 --> 0:00:10.320
<v Speaker 2>So here's the latest this morning, Apollo noting a lack

0:00:10.360 --> 0:00:13.280
<v Speaker 2>of profit margin gains due to AI outside of the

0:00:13.360 --> 0:00:16.680
<v Speaker 2>tech sector. Torston's slock of Apollo, writing, there's a mismatch

0:00:16.720 --> 0:00:20.480
<v Speaker 2>between current earnings expectations and the actual time firms need

0:00:20.840 --> 0:00:24.279
<v Speaker 2>to generate ROI on AI investments, and it could have

0:00:24.400 --> 0:00:29.080
<v Speaker 2>significant implications for many AI company valuations. Towston joins us

0:00:29.080 --> 0:00:30.680
<v Speaker 2>now for more. Torston, good morning, good to see you.

0:00:31.120 --> 0:00:32.600
<v Speaker 2>Let's build on that quote. What are you track in

0:00:32.680 --> 0:00:33.560
<v Speaker 2>right now? What do you think?

0:00:33.720 --> 0:00:36.280
<v Speaker 3>Well, what's really really important to this discussion is, of

0:00:36.320 --> 0:00:39.520
<v Speaker 3>course profit margers have been phenomenal in the Magnificent seven,

0:00:39.760 --> 0:00:41.559
<v Speaker 3>but what really is critical is that now we need

0:00:41.600 --> 0:00:44.360
<v Speaker 3>to see profit marners grow up outside the Magnificent seven.

0:00:44.400 --> 0:00:45.720
<v Speaker 3>In other words, what's going on with this in p.

0:00:45.720 --> 0:00:48.879
<v Speaker 3>Four ninety three becomes very very critical because at this

0:00:48.960 --> 0:00:50.199
<v Speaker 3>point profit margers is in P.

0:00:50.240 --> 0:00:52.199
<v Speaker 4>Four ninety three have just not gone up.

0:00:52.400 --> 0:00:55.040
<v Speaker 3>So therefore, one very important concusion, and one very important

0:00:55.040 --> 0:00:58.080
<v Speaker 3>place to look for signs of AI beginning to have

0:00:58.120 --> 0:01:00.720
<v Speaker 3>an impact is to look at what's going on in

0:01:00.760 --> 0:01:04.160
<v Speaker 3>earnings growth, profit margins, and overall the health of this

0:01:04.280 --> 0:01:06.240
<v Speaker 3>in p. Four ninety three, as a result of the

0:01:06.319 --> 0:01:08.400
<v Speaker 3>technological improvements we're seeing at the moment.

0:01:08.160 --> 0:01:10.319
<v Speaker 2>Mats are also making a call potentially as well that

0:01:10.400 --> 0:01:13.640
<v Speaker 2>the ROI on selling access capacity might be hard in

0:01:13.840 --> 0:01:17.160
<v Speaker 2>using it internally. Does that reinforce some of this message

0:01:17.160 --> 0:01:17.360
<v Speaker 2>for you?

0:01:17.840 --> 0:01:19.880
<v Speaker 3>Well, The issue, of course is that there is a

0:01:19.959 --> 0:01:22.160
<v Speaker 3>huge of course built out of capacity and compute, and

0:01:22.200 --> 0:01:25.480
<v Speaker 3>there will literally be unlimited demand for compute. The question

0:01:25.600 --> 0:01:27.800
<v Speaker 3>is just at what price and who will the buias

0:01:27.840 --> 0:01:30.320
<v Speaker 3>be and where is that capacity coming from? And the

0:01:30.360 --> 0:01:34.160
<v Speaker 3>big picture still remains that AI is a very revolutionary technology.

0:01:34.319 --> 0:01:35.360
<v Speaker 4>Everyone agrees on that.

0:01:35.600 --> 0:01:37.360
<v Speaker 3>But the key question now is how long time is

0:01:37.400 --> 0:01:40.000
<v Speaker 3>it going to take before this shows up, especially in

0:01:40.040 --> 0:01:43.480
<v Speaker 3>profit margins outside the Magnificent seven, Because if there's in p.

0:01:43.560 --> 0:01:45.920
<v Speaker 3>Four ninety three, let's say that it takes several years

0:01:45.920 --> 0:01:48.360
<v Speaker 3>before profit margins begin to go up. The question is

0:01:48.360 --> 0:01:52.200
<v Speaker 3>whether the implicit earnings assumptions in the Magnificent seven are

0:01:52.240 --> 0:01:55.440
<v Speaker 3>too high or too fast relative to what's actually going

0:01:55.440 --> 0:01:58.640
<v Speaker 3>to happen, so that mismatch between are we going to

0:01:58.640 --> 0:02:02.160
<v Speaker 3>see profit margins earnings growth go up? Outside this the

0:02:02.240 --> 0:02:04.920
<v Speaker 3>Magnificent seven. Is that going to come slower? Is it

0:02:04.960 --> 0:02:07.720
<v Speaker 3>going to come faster? Is absolutely critical for a conversation

0:02:07.800 --> 0:02:10.480
<v Speaker 3>about what should the value be of the Magnificent seven.

0:02:10.520 --> 0:02:13.920
<v Speaker 1>Today, in exactly one week's time, as John was reminding us,

0:02:13.919 --> 0:02:15.880
<v Speaker 1>the big banks begin reporting earnings and we have the

0:02:15.960 --> 0:02:18.040
<v Speaker 1>kickoff of the earning season, there's going to be a

0:02:18.040 --> 0:02:20.320
<v Speaker 1>lot of discussion on AI and what it means for

0:02:20.600 --> 0:02:23.600
<v Speaker 1>jobs in the banking sector. How do you parse through

0:02:23.720 --> 0:02:26.840
<v Speaker 1>what the companies say to really understand what it means

0:02:26.880 --> 0:02:28.800
<v Speaker 1>in terms of whether they're going to cut jobs or not.

0:02:29.080 --> 0:02:31.960
<v Speaker 3>What's really challenging about this is to talk about AI

0:02:32.120 --> 0:02:34.800
<v Speaker 3>exposure because there's a lot of different studies already that

0:02:34.919 --> 0:02:38.400
<v Speaker 3>look at what is the exposure meaning AI exposure in

0:02:38.440 --> 0:02:42.880
<v Speaker 3>different occupations. Two these margets, These two studies fall into

0:02:42.880 --> 0:02:46.359
<v Speaker 3>different buckets of one saying what is the actual exposure

0:02:46.400 --> 0:02:49.160
<v Speaker 3>in terms of actual AI usage? So this is trying

0:02:49.200 --> 0:02:51.800
<v Speaker 3>to measure, say, what are people using clode for what

0:02:52.000 --> 0:02:55.560
<v Speaker 3>tasks and what requests did they get and therefore measuring

0:02:55.560 --> 0:02:59.359
<v Speaker 3>and quantifying the actual usage of AI. Another bucket is

0:02:59.400 --> 0:03:03.080
<v Speaker 3>studies that look at, well, let's theoretically assume what is

0:03:03.120 --> 0:03:06.080
<v Speaker 3>the economistic exposure to AI and then try to match

0:03:06.120 --> 0:03:08.720
<v Speaker 3>that with occupations and figure out what is employment in

0:03:08.760 --> 0:03:11.239
<v Speaker 3>those sexes. And those studies, of course, are what you

0:03:11.360 --> 0:03:13.920
<v Speaker 3>call more theoretical. So the challenge is that there is

0:03:13.960 --> 0:03:17.000
<v Speaker 3>not an agreement about what does AI exposure mean, So

0:03:17.040 --> 0:03:19.120
<v Speaker 3>that's raising all these questions around, well, what does it

0:03:19.160 --> 0:03:20.959
<v Speaker 3>even mean when we say that a certain part of

0:03:21.000 --> 0:03:23.360
<v Speaker 3>the economy is exposed to AI, because that's just not

0:03:23.440 --> 0:03:25.120
<v Speaker 3>at this point in the studies that look at this

0:03:25.480 --> 0:03:28.240
<v Speaker 3>a really good way to quantify what AI exposure means.

0:03:28.280 --> 0:03:30.320
<v Speaker 3>And that means also for the financials, that it means

0:03:30.320 --> 0:03:33.080
<v Speaker 3>for legal services, that means for consultants that we're having

0:03:33.080 --> 0:03:36.280
<v Speaker 3>some challenges figuring out how do we even quantify what

0:03:36.400 --> 0:03:37.920
<v Speaker 3>is the impact of AI. We all know that it's

0:03:37.960 --> 0:03:40.000
<v Speaker 3>going to make a big difference, but to speak with

0:03:40.040 --> 0:03:43.160
<v Speaker 3>which this difference comes along? How many people is impacting today,

0:03:43.320 --> 0:03:46.160
<v Speaker 3>next month, next year. It becomes absolutely critical when you

0:03:46.200 --> 0:03:48.960
<v Speaker 3>again think about that company valuations today are the net

0:03:48.960 --> 0:03:51.520
<v Speaker 3>present value of the cashflows that these companies get in

0:03:51.560 --> 0:03:51.920
<v Speaker 3>the future.

0:03:52.000 --> 0:03:54.760
<v Speaker 1>We're also hearing a changing story, a changing narrative from

0:03:54.800 --> 0:03:57.400
<v Speaker 1>AI companies themselves. There's a Wall Street Journal story about

0:03:57.400 --> 0:03:59.960
<v Speaker 1>how the narrative has shifted, at least from open AI

0:04:00.320 --> 0:04:03.360
<v Speaker 1>and anthropic from these doomsday scenarios to a future where

0:04:03.400 --> 0:04:05.800
<v Speaker 1>workers actually keep their jobs but just do it better

0:04:05.840 --> 0:04:06.440
<v Speaker 1>thanks to AI.

0:04:06.560 --> 0:04:07.360
<v Speaker 4>What does that tell you?

0:04:07.480 --> 0:04:10.120
<v Speaker 3>Yeah, and RAMP had a really interesting study over the

0:04:10.200 --> 0:04:12.520
<v Speaker 3>last week that they put out where they basically look

0:04:12.600 --> 0:04:15.920
<v Speaker 3>at what has been the cost and the spending on

0:04:16.080 --> 0:04:18.920
<v Speaker 3>AI among different companies and what they did that they

0:04:18.960 --> 0:04:21.240
<v Speaker 3>looked at, well, what was the job growth in those

0:04:21.240 --> 0:04:23.680
<v Speaker 3>companies that had more spending on AI? And they did

0:04:23.680 --> 0:04:26.960
<v Speaker 3>indeed find that more spending on AI, it actually resulted

0:04:27.040 --> 0:04:29.280
<v Speaker 3>in more job growth. So one way of looking at

0:04:29.279 --> 0:04:31.880
<v Speaker 3>this is one dimension of saying, what's the actual spending

0:04:32.120 --> 0:04:36.120
<v Speaker 3>relative to the more theoretical matching of occupations with what's

0:04:36.160 --> 0:04:37.920
<v Speaker 3>been going on with job growth in those sects that

0:04:37.960 --> 0:04:39.040
<v Speaker 3>also have been spending on AI.

0:04:39.160 --> 0:04:41.200
<v Speaker 2>Tostened to your points, put this all together, this is

0:04:41.240 --> 0:04:44.600
<v Speaker 2>a potential risk of valuations, which is an obvious market risk.

0:04:44.720 --> 0:04:47.360
<v Speaker 2>Is it a macro risk as well? In Central Portugal,

0:04:47.400 --> 0:04:50.360
<v Speaker 2>I believe you were there big conversation about this. We've

0:04:50.400 --> 0:04:52.479
<v Speaker 2>got a story in the market right now that feels

0:04:52.560 --> 0:04:55.000
<v Speaker 2>like one trade and increasingly a story in the economy.

0:04:55.040 --> 0:04:57.080
<v Speaker 2>The Fiells lie with firing on one engine, how much

0:04:57.200 --> 0:04:58.160
<v Speaker 2>is it one and the other?

0:04:58.400 --> 0:05:00.800
<v Speaker 3>Absolutely, it's actually three different things. First of all, it's

0:05:00.839 --> 0:05:03.599
<v Speaker 3>of course everywhere in markets because the concentration to the

0:05:03.640 --> 0:05:05.039
<v Speaker 3>AI story is so strong.

0:05:05.080 --> 0:05:05.640
<v Speaker 4>Think about it.

0:05:05.760 --> 0:05:08.000
<v Speaker 3>For the last fifteen years, the main lesson in finance

0:05:08.080 --> 0:05:10.880
<v Speaker 3>is factor investing, and now we're staring at one factor

0:05:10.920 --> 0:05:14.080
<v Speaker 3>that's driving all financial markets. In equities, the concentration is

0:05:14.120 --> 0:05:16.560
<v Speaker 3>very high in AI. You also look at IG issuance,

0:05:16.640 --> 0:05:18.680
<v Speaker 3>high yield issues, even venture capusal.

0:05:18.880 --> 0:05:20.839
<v Speaker 4>The concentration in AI is very very strong.

0:05:20.880 --> 0:05:23.520
<v Speaker 3>So in markets, let's just agree the AI exposure or

0:05:23.520 --> 0:05:26.040
<v Speaker 3>the AI factor plays a very very critical role at

0:05:26.040 --> 0:05:28.560
<v Speaker 3>the moment when it comes to the economy. Also, if

0:05:28.560 --> 0:05:31.200
<v Speaker 3>you look at actual spending on data centers and energy,

0:05:31.440 --> 0:05:33.520
<v Speaker 3>if you add up what that contributes to GDP this

0:05:33.600 --> 0:05:36.080
<v Speaker 3>year is roughly about zero point seven percent out of

0:05:36.120 --> 0:05:37.720
<v Speaker 3>a two percent GDP growth this year.

0:05:37.880 --> 0:05:39.360
<v Speaker 4>If you add about zero point three.

0:05:39.240 --> 0:05:41.440
<v Speaker 3>Coming from the wealth effect because of high stock prices,

0:05:41.560 --> 0:05:44.560
<v Speaker 3>that's also very important. And finally, let's also not forget

0:05:44.720 --> 0:05:48.320
<v Speaker 3>that the hyperscalers are issuing so much IG corporate debt

0:05:48.560 --> 0:05:52.119
<v Speaker 3>that this is crowding out demand for US treasuries because

0:05:52.120 --> 0:05:54.240
<v Speaker 3>if you are bunt manager and investment great credit, you

0:05:54.240 --> 0:05:57.360
<v Speaker 3>could either buy sovereigns US treasuries, you can buy financials.

0:05:57.360 --> 0:05:58.880
<v Speaker 3>That's what you've been doing for a long time. But

0:05:58.920 --> 0:06:01.479
<v Speaker 3>now you can also buy seven hundred billion hyperscalers. So

0:06:01.560 --> 0:06:03.560
<v Speaker 3>it's not only that it has an impact on markets

0:06:03.600 --> 0:06:06.040
<v Speaker 3>overall and has an impact on GDP, but it actually

0:06:06.080 --> 0:06:06.600
<v Speaker 3>also has an.

0:06:06.560 --> 0:06:08.000
<v Speaker 4>Impact on demain for treasury.

0:06:08.040 --> 0:06:10.520
<v Speaker 3>So yes, the conversation and the panel Amazon in CenTra

0:06:10.800 --> 0:06:14.080
<v Speaker 3>was exactly around this risk that AI is indeed now

0:06:14.120 --> 0:06:16.520
<v Speaker 3>and more and more prominent factor basically not only the

0:06:16.520 --> 0:06:18.360
<v Speaker 3>economy but also in financial markets.

0:06:18.360 --> 0:06:20.159
<v Speaker 2>What's the consensus on how to manage that risk?

0:06:20.640 --> 0:06:23.279
<v Speaker 3>Well, the challenge is page one in your finance textbook.

0:06:23.360 --> 0:06:25.200
<v Speaker 3>If there's one factor you're trying to avoid is to

0:06:25.279 --> 0:06:27.800
<v Speaker 3>try to pick another factor. But the question is what

0:06:27.920 --> 0:06:30.800
<v Speaker 3>is that other factor? Momentum is a growth is a value?

0:06:31.000 --> 0:06:34.960
<v Speaker 3>Value has some opportunities because it's not growth. So that's

0:06:34.960 --> 0:06:37.920
<v Speaker 3>why ways of looking at parts of the private markets,

0:06:37.920 --> 0:06:41.240
<v Speaker 3>public markets, that is value investing is indeed one place

0:06:41.279 --> 0:06:43.320
<v Speaker 3>to hide, but it has to be value investing. That's

0:06:43.360 --> 0:06:46.119
<v Speaker 3>indeed protecting you against the downside risks that come along

0:06:46.400 --> 0:06:48.840
<v Speaker 3>if Ai does not deliver in the.

0:06:48.839 --> 0:06:52.080
<v Speaker 2>End, Towston's Slock of Apollo toaston fantastic no great rate

0:06:52.320 --> 0:06:54.240
<v Speaker 2>in the last week orse, so I appreciate it. Towson'slock

0:06:54.240 --> 0:06:56.080
<v Speaker 2>of Apollo. There On, Ai and tag