Bloomberg Audio Studios, Podcasts, radio News.
This is the Bloomberg Surveillance Podcast. I'm Jonathan Ferrow, along with Lisa Bromwitz and Amerie Hordert. Join us each day for insight from the best in markets, economics, and geopolitics from our global headquarters in New York City. We are live on Bloomberg Television weekday mornings from six to nine am Eastern. Subscribe to the podcast on Apple, Spotify or anywhere else you listen, and as always on the Bloomberg
Terminal and the Bloomberg Business app. We begin this out were Stock's pushing kaya following a boost from Micron's blowout earnings report. Sebastian page of two, Row price writing AI is an expanding theme from GPUs to other bottlenecks along the data center supply chain, as buzz light Years said to infinity and beyond. Sebastian joins us now for more. So, welcome back to the program, Buddy. Are you just saying, stay on grip tie, keep rounding this ball?
John? When I wrote my notes, I didn't think you'd pick up on the buzz light year quote. I'm going to be honest. We have an expending trade from bottlenecks to more and more bottlenecks. You look at cooling, at electrification, at aerospace, at gas turbines, We're invested in those bottlenecks. We think it's a trade that has legs. Look the AI build out. A lot of people mentioned those forecasts on your show, but we're looking at three trillion by
twenty twenty eight. So it's a trade and it's going to continue to go. In my mind, the key question right now is when is this going to show in AI users' margins. Now we have three hundred analysts on our platform. I read their notes, but now I use AI to read their notes, and I ask AI, where are we seeing companies in our own proprietary research that
are increasing their margins because they're using AI. Now I've seen estimates out there John that are I think exaggerated, but anecdotally we're starting to see this as well in the users of AI. This will be the next leg. So I think that technology revolution will continue. I'm conscious as I say this I sound a little bit like Dan Ives, but yeah, this we're long US large cap growth stocks.
So I'll push back just a little bit and I'll give you the alternative view on things step where the economics of all of this get a little bit complicated. So the cost of building capacity is increasing. We can see that evidenced by some of the margins at some of these companies at a time when the end user demand is getting a little bit fragile with regards to pricing goverall, and they're rationally spending on AI. How's that going to make sense going forward? What we need to rationalize here?
Look, as long as the products keep improving, and they are there is tremendous demand for it, it looks to me like we're still scratching the surface on demand and usage of AI. And by the way, John, I completely take the point that this spending is gigantic and free cash flows are coming down really fast for a lot of the hyper scalers. But you know, if you think about it, the price earnings on the Russell one thousand growth, we were in the thirty range. We peaked at thirty
several times over the last five years. John, We're at twenty two. So there is you know, trust the market. There is an adjustment here that's been made. You know, this is this is in the bottom for Russell one thousand growth valuation. This is in the bottom thirty percent of the last ten year range. You can also play AI bottlenecks, by the way, in other areas. You can also play them in small and MidCap, which we're long too. We have this barbell between US larscap growth and small
and MidCap. We're talking about the golden component. Once you have a component that becomes a part just a bottleneck, a part of that electrification, power supply, data center, whatever it is, your company stock goes vertical. Also, merging markets, A lot of people mention this on surveillance. Emerging markets are becoming part of that trade. So yeah, I mean it's a position you need to balance. You need to
diversify the risk in the portfolio. For example, I still think there's inflation risk in the portfolio, and at some point that becomes part of the supplying the energy to AI, that becomes part of the picture.
Well, pick up on that just one second, SEB.
You're talking about trying to race to find the next bottleneck where you can see stocks go vertical, akin to what we're seeing with Micron this morning. I just wonder if people truly are understanding or if it's a healthy technical They just have a rolling ball of cash. On one hand, people are chasing who has pricing power. On the other hand, Google shares down eight percent this month. Microsoft poised for its worst monthly loss going back to
twenty and eight of nineteen percent losses. The payers of this bill are getting penal At what point is that a warning side?
I mean, it is absolutely, Lisa. You have winners and losers. And this is why skilled active management and fundamental research is working here. You're going to have software companies that are going to be disrupted, but you're going to have companies that, I think increasingly the next leg of this is along the bottlenecks to eventually the handoff to the users that are going to create real efficiency. I see it in our company. We're creating efficiency, We're improving our processes.
It's really that that's going to be the next leg, and that's what is going to keep it going. And again I go back to the fact that it's not you know, a lot of this increase in leverage and borrowing and decrease in free cash flow margins, a lot of it has been adjusted in the price twenty two relative to the thirty range. In prior episodes over the last five years, I think is pretty good for Russell one thousand growth.
Se could you justify the investment and further AI, if the borrowing costs were higher in the face of inflation, in the face of some of this acceleration that we're seeing in other components and price components that everyone's trying to chase.
Well, the reality is that this three trillion is spending. A lot of it is generating, generated by companies that still have a ton of cash. I don't like the comparisons to the Duck Humbubble, but we're nowhere near the
level of corporate leverage and aggregate that we had. So yes, there is some borrowing, but if you look at the metrics by historical standards, it's not a very high level of leverage by historical standards, And there is tremendous capital that's available, as we're seeing with IPOs and you know those issues. I think there is there is demand for the end product at the end of the day, that's what really matters. People want to use it. They're billions
of users already, they're just scratching the surface. I'm doing roundtables with investors and I realize a lot of them, my peers, a lot of them are just starting. It makes me happy because I feel like we're really ahead and how we're using agents and so on in our investment process. But also it makes me think that as long as the demand for the end product is there, people are willing to pay for it. We'll go through
ups and downs. We'll go through episodes where tokens get more expensive and people go to cheaper models and so on, But the trend is our friend.
Here.
Three trillion spending by twenty eight. Again, Lisa, I didn't think I come in this morning and sound like dan Ives, but we do have a long position in US large gap growth, and we do believe in the AI bottlenecks and the expansion of that trade.
You also are concerned that even with oil prices becoming lower, that you actually think there's still a lot of inflation baked into the system. Is a fed potentially having to maybe hike going to be an issue for where you see the market going and your dan Ives bullishness this morning.
Yeah, Look, I'm Marie. If we like twenty five base points fifty base points of hikes, I don't think that's a big issue for markets. I do think generally though, inflation right now is underestimated as a risk. You had the Torsten Slock quote earlier, I would argue that he's right. I would agree with that, Yes, oil prices are coming back down really quickly. My view is that there's spent up inflation. There are lots of lags in inflation data
and in inflation itself. You have from fertilizer to food, from freight and insurance to transportation, to industrial production to the final goods, from fiscal stimulus to demand. There's lots of lags that are working their way through. By the way, the last three months on the rolling basis, the headline CPI is up eight percent. If you annualize it, you're not supposed to analyze three months. You know it's noisy data, but still eight percent. But the shelter for the last
three months is four point eight percent. There's also a lag de fact here with prior rent increases. Now, put all of this in context of the fact that the break even for the one year horizon is one point seven percent and the one year inflation swap is two point two percent. So my mind is still an underestimated risk especially for the next three to six months. In the long run, yeah, will normalize, but I think the
market is underpricing the inflation risk. I guess we have a print this morning, We'll see I watch surveillance often with a twelve or twenty four hour lag, and I always I'm mindful that some people come in and make forecasts a few minutes before the data release. I'm not going to make that. I'm I'm not going to do that this morning. It's always interesting to listen to it
after the fact when you know what happened. But as a general trend, I'm not talking necessarily PC maybe four percent this morning, but as a general trend, the inflation here is underestimated by the market based on the pricing you've seen in break evens and inflation swap.
Stay with us, Mulblinberg, Savana's coming up off to this, Micron's share searching on its blockbus to forecast. Hey, Terry of City Research right in the health of the AI trade will continue to be defined by the irresistible force of enterprise demand meeting the immovable object of inference capacity. Heith joins us now for more. Hath good morning, good to see it.
Good morning.
Let's pick up on that enterprise demand. Just how robust is that right now?
Look, it's incredibly robust. We'll get Q two numbers in a few weeks. But if you look at what we've seen over the last the last few quarters, we've gone from twenty eight percent year over your growth in Q one of last year to thirty three percent in Q two, to fifty three percent in Q three, to ninety one percent to one hundred and forty one percent in Q
one of this year. Like that kind of bend in the S curve that we're that we're seeing here and we believe it's going to accelerate again in Q two. That's the best proxy for what's happening in enterprise demand. And anybody who works at a Fortune five hundred type company that is pushing this into products could can tell you they're seeing it in real time.
There is a rush to build out capacity to make all this demand. Some companies are benefiting both than others Microme with eighty five percent margins. Other companies that are doing a lot of the spending right now being punished. The stocks are being punished. This is a thing we keep bringing up that changes the trajectory for spending.
Look, no, I mean not anytime soon, because the companies are getting real returns on this. The market right now is not giving the companies that are spending on this, the hyperscalers the kind of credit that they normally would for a couple of different reasons, but the biggest one is there is this gap between when you have to start spending and building all of this out when you
actually start seeing the returns on it. If you look at cash return on cash invested, which we believe is sort of the best gauge of the returns that they're actually generating on this, hyper scalers are generating twenty nine percent returns on this investment, very stable. It was twenty nine percent in Q one, is twenty nine percent in Q four. If you have the ability to borrow at six percent or to issue equity and get a twenty nine percent return on it, you're supposed to do that
regardless of what your stock price is doing. And while the stock prices aren't doing what Micron stock price is doing, they're hanging in there enough that investors haven't rejected this.
Right.
Google issued eighty five billion dollars worth of equities the market didn't really blink.
Well, just to build on though what John is talking about, maybe people are still lending and they're.
Still able to raise cash.
Nonetheless, it's getting a little bit hairy for the likes of Microsoft down eighteen percent for this worst monthly performance going back to two thousand and eight. You see Oracle down more than thirty percent. Are we starting to see the beginnings of investor pushback that maybe hasn't put teeth in this trade by withdrawing cash, but certainly has put a little bit of notice around that capital.
Well, the thing that we have to remember, particularly when it comes to Microsoft and Oracle, is those are both software companies. Look at everything else in software and what those stocks have done. The market's not responding negatively to the acceleration and growth that you're seeing at Azure, the returns you're seeing at Azure. It's responding to the other seventy percent of Microsoft's business that is software. It's responding to the biggest part of Oracle's business, which is software.
And so I think that's that, if anything, should encourage more investment in this area, particularly as we start to see the acceleration in Azure revenues through through the end of this year, which we believe the market will respond positively to.
We've seen this rolling ball of cash, and as we heard from Sebastian Page earlier, everyone's looking for the choke points. The next place where you're going to see things just sort of shoot exponentially higher because of this insatiable demand and a limited supply. After memory, what's next, I mean, do you have your eye on kind of those choke points.
I mean we do.
I mean the choke points sort of exist throughout the supply chain right now, Memories obviously the biggest one. Anthropic told you that very loudly a few weeks ago when they raised their most recent round. They have the opportunity to choose strategic partners. They could have picked anyone they wanted.
They chose three, Samsung, Micron, and Heinex, and so like if anyone who's surprised by what's happening with Micron today and what's been happening with high bandwidth memory didn't pay attention to what Anthropic told them, the next point after that is it's at interconnect, it's in storage, it's in labor in a very big way, and of course it's in power. We've been talking about that for so long.
It's also because of what you see with Micron. It's very much in semicap equipment because we do have to make more memory, we do have to make more chips, We need more compute, we need more storage and memory.
Do you think some of these companies along the supply chain we could potentially see these issues emerged later in the year or next year as they build this out. Are undervalued right now?
Well?
I mean, so we've been saying this, like from I started at City a year ago. The first thing that we said is you want to be where the model necks are, and that has continued to be the case, and we sort of stuck with that. This is where pricing goes. You were talking about sort of the underperformance of the hyperscalers. You want to be where people are spending money, not necessarily with the people who are spending the money, and so that's going to continue to be
the case. We would argue it's not even so much undervalued as just estimates are too low, Consensus numbers aren't high enough. Because the amount of infrastructure that is going to get built, the amount of pricing increases that you're going to see that, you know, irresistible demand, immovable supply. The only thing that breaks in that is price. You saw in the Micron number eighty five percent gross margins. Memory companies are not supposed to have eighty five percent
gross margins. Eventually, that will settle out because that's what cyclicals do. But we're a long way away from that happening.
You said that one of the choke points was in labor. Are you talking about Jonas Adler Are you talking about Dave who constructs some of what we're seeing with the hyperscaler buildings.
So it's certainly both. There is not a lot of talent out there that has the capability of advancing these models the way that level of sort of the Jonas Addlers of the world do.
I have no idea, but.
I mean, if you look at some of the stuff that was announced last summer, right when Meta was doing these deals, right, Anthropic is doing these deals now, Meta was doing them last summer, it was six and in some cases, you know.
I mean it was it was.
Very large sort of you know, in some cases billion dollar kind of pay packages over multi multi years. And when you look at sort of the acquisition hires that many companies have done. It falls into that hundreds of millions and beyond kind of level. That said, what I was referring to is very much it's the electricians, the plumbers, the construction people that actually have to build these data centers. There are very few people in the world that have the skills to be able to build at the level
that these data centers require. And those people are in massive demand right now, and deservedly they're getting paid a lot more to stay with us.
Mult Bloomberg surveyance coming up after this check Kevin Walsh making changes at the Central Bank limited forward guidance task forces and putting even more way on signals from financial markets. The former New York Fed President Bill do be Wang in writing this a fresh look is appropriate, but this needs to be done with greater care than w WASSH is shown to date. Bill joins us now for more. Bill,
welcome to the show. It's nice to reflect on these kind of things with you, So thanks for being with us. What are you reflecting on right now? What are you concerned about? Is this comes together?
Well?
What I'm concerned about is that I'm perfectly fine with it getting rid of four guns. But what I'm concerned about is worsh is not disclosing what is monetary policy reaction functions. In other words, how would the FED adjust policy as the economic environment changes. And I think that's a big mistake and the idea that you can rely on financial markets to tell you what you should do in terms of monetary policy, that won't work because markets
priced to what they think the FED will do. So if the Fed's looking at the markets and the markets are looking at the FED, how is policy actually set. It's really important in the United States that markets understand how the Feds could react, because monetary policy works mainly through financial conditions, the effect of short term rates on
bond stocks, credit spreads, the dollar. So if the markets don't understand what the FED is doing, uh, those markets aren't going to be priced appropriately relative to what the FED is actually going to do. And that's going to both slow down the transmission of mandre policy to the real economy, but it's also going to make it less, you know, less less efficient.
So it seems to me like you've.
Got to provide more guidance about how you're actually going to react if things evolve differently than you anticipate.
To your point, Bet, I think you're picking up on attention. We are witnessed in the news conference he seemed to find difficult to separate forward guidance from communicating articulating a reaction function. Where do you think that came from? Why is that difficult? Do you think it's straightforward?
I think it's straightforward.
I mean, I think Ford guidance is really basically saying, here's what we expect to do next. The madre policy reaction function isn't about what we're going to do next. It's about how we would react to different sorts of incoming information.
So I think he needs to distinguish between those two things.
Right now, I don't really understand how much how he expects to set mandreate policy. The job of mandre policy is the fedest job, not the market job, and so the FEDS needs to do its job.
Part of his argument, perhaps just being generous, might be that their task forces for that, and essentially they have to understand what data they're looking for at before they understand exactly how they should react.
Do you buy that argument, Well, obviously it makes sense to take a fresh look at what data is available, but at the end of the day, these task forces have a finite life. I mean, the FED has basically got to conduct mandre policy not just over the next six months, but over the next number of years. So I think the reliance on task forces maybe it buys him some time to think about what he really wants
to do. But at the end of the day, these task forces are not a replacement for the FED communicating about how it's going to act as economic conditions change.
Right now, BO, when you take a look at the inflation ry, when you take a look at the oil input, but also the broadening as we're seeing it in core PCEE, which we're going to get in about nine minutes time, do you think it is appropriate for them to high at least once, if not twice or three times this year.
Well, I think that the case for monetary policy becoming a bit tighter is pretty compelling to mebe for two reasons. Number One, we've been at this level of rates or higher for three years and the economy is still at a four point three percent unemployment rates, So what's the evidence that monetary policy is restrictive? And second, financial conditions
are really accommodative. The board has a model of financial conditions, and right now it shows the impulse to growth over the next year is over one percent positive on GDP growth. That's the highest since late two thousand and one early twenty twenty two. So I think that easy financial conditions, no evidence that monetary policy is restrictive at the time that you've missed your target for more than five years does create a strong argument for tightening montary policy.
I'm aligned with you know, Alberto Missal and Lori.
Logan, But what I think doesn't really matter is really what Kevin Warish is ultimately going to do. One thing is going to help a little bit in his you know, giving them a little bit more time. Is this is probably the really last bad headline inflation report we're going to have for a while, because with the decline and l prices, when we get the inflation data for June, headline inflation at both the CPI and PC level is going to decline quite significantly.
But what about the housing market. Isn't it still fair to say who we have is highly restrictive for housing.
Well, I think housing is not doing well for a couple of reasons. The main reason though, is lack of demand because we don't have any labor force growth. So if you don't have labor force growth, you don't have household formation. If you don't have househole formation, you don't have a lot of demand for houses. There's also an affordability issue in terms of what level of income you have to achieve to be able to buy a house. But I don't think you know, you look at mortgage rates.
Our mortgage rates particularly high. I mean they're high relatives the last fifteen years or so. But if you go back prior to the Great Financial Crisis, you know, mortgage rates in the six and a half a percent range.
No one would view that as particularly high.
Well, it is high, though, if you're sitting on a three percent rate and fuse to move because then it doesn't offer up a lot more housing for everyone else, well.
There's a lock in effects that people are deciding not to move because they don't want to lose the advantage of those very low mortgage rates. But if they move, they're they're demanding another house. So I'm not sure that you know that really is going to have a big effect on housing affordability. A big, big question I think if for a house of affordability is what can you do about zoning, what can you do about land use
to basically allow more homes to actually be produced. Increasing the supply of housing would then weigh on housing prices, and that would make housing more affordable. But I don't think it's really a big interest rate problem, frankly.
But just a final thread, a few questions just on this institution. You've worked at it a long time, from your position at the New York FED, going down to Washington on so many occasions, how easy is it to change this institution?
Well, there's certainly a lot of inertia in the sense that do you have staff that are there going to be there a lot longer than Kevin Wosh is going to be there, for example. And the staff has a lot of pride in their work, and deservedly so because they have a tremendous amount of expertise. But you can
definitely move things if you have a better mousetrap. And when I went when I went down to the near FED after the first day, when the second day I was president of the r FEDI basically said my mantra, best idea wins. The best idea that people come up with should actually dominate, regardless of where it comes from. And I think the board staff, the staff of the
federals or banks really subscribe to that. So if Kevin Warrish and the task forces come up with better ideas, I think those best ideas will actually win.
This is the Bloomberg Survendment's podcast, bringing you the best in markets, economics, angiopolitics. You can watch the show live on Bloomberg TV weekday mornings from six am to nine am Eastern. Subscribe to the podcast on Apple, Spotify or anywhere else you listen, and as always, on the Bloomberg Terminal and the Bloomberg Business app