AI credit supply tests market capacity

0:00Hi, everybody. Welcome back to the Barclays Brief. It's Ronnie. It's July 27th, and I'm here in our studio with Dominique Toublan, our head of U.S. credit strategy. Dominique, your first time on the podcast. Welcome. Thank you. Hi, Ronnie. Thanks for having me. Great to have you here. And the timing is perfect because all of a sudden, equity investors are talking about credit spreads in tech. So we've talked a lot about AI investment, AI from every angle on this podcast. Now, all of a sudden, it's a huge topic for the investment-grade credit markets. Can you walk us through why that's happening and what's changing in credit markets?

0:36The basis of the AI revolution and how much CapEx is needed to make that happen. Companies have to borrow money to get there. They cannot generate enough free cash flow to get there. So borrowing money is what they're doing. And they are borrowing using every channel at their disposal. One of the biggest channels that is available to them is the public bond market. And so that has been growing enormously. year-to-date the AI ecosystem has issued about 20 % of the total debt issued in corporate bond markets on the US side. That's really, really large. And that's why people are paying attention.

1:06Usually when an equity investor starts to pay attention is that there are some worries around the ecosystem. Yeah, I agree with that. So now it feels like we're dealing with a capacity issue, possibly a quality issue. How do you think these hyperscalers are going to continue to influence the IG market? And is this about capacity or quality? So exactly right. Capacity is the issue here is how much they need to issue. If you look at the most recent earnings, top nine was good in the most recent hyperscalers that I've reported. So that doesn't seem to be the problem here. In credit fundamentals, the group altogether is high singularity.

1:42So again, doesn't seem to be dramatic anything there. It's really how much is coming and the surprise. So for instance, it's one that came, you know, surprised the market a little bit a month or so ago and the market say, oh, I have to make space for this. I didn't expect that one to come. And the bite size was quite large compared to what people expected. So they had to pay a higher new issue concession. They had to pay for it a little bit. That surprise cost them some money. So now the question is how much is still to come. So I don't think spreads are pricing so much fundamental, but it's more about what is next to come.

2:14So now if you invest in You know, in hyperscalers, at that, it's a little bit like if you're sitting down at a three-star Michelin restaurant and you don't know how many courses are going to come. So the food's going to be good, but you're going to have to pace yourself. That's the situation now. How much more is to come given how much more CapEx they are spending and how much more CapEx they say they will spend going forward? Okay, so it's$220 billion in U.S. hyperscaler CapEx investment grade issuance at this point year to date. How big do you think that's going to get from here? For the full year, we think, you know, high$200,$285 across different currencies for the group.

2:49But that's probably going to continue to grow into next year. And that's something, again, something to highlight in terms of they are using all the channels possible. So they're using equities, they're using converts, ABS, CMBS, private credit, public credit in dollar, euros, CAD, Swiss francs, whatever, you name it. They're using everything that is available to them because the numbers are so large. Right. And those large numbers feel like they're driving a little bit of a buyer strike, a little bit of reluctance to step in. What's going on? That's right. People are, I think, just trying to look from the side a little bit and say, I need to understand better what's going on for me to step in.

3:26We have more hyperscalers reporting in the next few days. Usually the issue is right after that. So people are anxiously waiting to see in the very short term what's going on. So altogether, the second thing that investors are doing is looking across channels, right? Historically, the ABS and public bond guys, they didn't necessarily talk about risk together, maybe outside of macro risk. Now they have to break down the silos and talk to each other. So more clients I talk to are going to that model where we want to look across the whole board because we are exposed ultimately to the same company across all these channels.

4:00Right. And just very recently, we've started to see less of that. And we've seen this spread widening happening. What do you think this means for the broader AI narrative. Is this a real signal of fatigue? Is this something that equity investors should be staying very close to given what debt markets tend to mean or signals they send for our world? I think they should, right? Again, it's the incredible CapEx numbers that these companies are pushing out means that they have to finance it in some ways, right? So you will care whether they generate enough free cash flow or not to get to these numbers.

4:30And ultimately that they have enough ROI on all the investment they're making. You will definitely care about that at some point. The going forward is really for these companies to think about how do they finance this? And that might lead to financing that is kind of off balance sheet. Is it some cheap financing and you do leases on it? Is it data centers and you list them, et cetera, et cetera. So that's going to be, I think, investors, issuers going forward to try to be imaginative, try to construct new ways to finance this. Because if there's fatigue in the public market, which is the largest and most liquid one, you will need to find other outlets.

5:03Right. So there's a lot of complexity here around the financing. You're starting to see spreads widen. What else tells you that investors are becoming cautious? One neat way to look at this for us is we issue all the time, right? So you can look at this data and see what's going on. If you look at the overall market, we move the AI ecosystem for a second. The new issuance has done very well. A few metrics we look at is new issue concession, right? How cheaper you issue compared to existing bond? How much subscription there is? Is there appetite in the markets for it? And then what is your performance right after?

5:34Usually you do well, right? You issue cheap. Well, for the hyperscalers, as you were mentioning earlier, they were doing quite well until the second quarter. And since the second quarter, we've seen things reverse, where new issue concessions are high, oversubscriptions are low, and new issue performance is not great. So that's one place where it's the huge liquidity instant for these companies, and you can see directly what the market is doing. So that primary issuance is showing you that there is some friction, maybe some fatigue there. The second place you can see is just in Relval. Just to give you a couple of ideas.

6:04One, you take the best hyperscalers. They're trading wider than the index by about 10 basis points. A year ago, they were trading 50 tight. Wow. If you look at the very long end, where they've issued so much, because their assets on the data center, et cetera, side are long-dated. They like to have long-dated liabilities. The long end of all the hyperscalers trades on top of the double B index. So a huge difference in ratings, same spread. There's a difference in duration, don't get me wrong, but still, I don't think I've seen that before. So that feels like all the different channels of supply colliding with each other and creating that repricing.

6:39Can you just get a little more specific on what these other channels are and where else this financing is coming from? Sure. So you can see, again, across the board in different fixed income assets perspective, private credit, for instance, obviously has been a big one. There's some new deals that are being circulated. Well, they tend to come at a discount, so they're going to compete directly with the public bond market. But it's also there in CMBS, in ABS. If you go away from the fixed income market, you see them as well on converts and equities. So again, all these channels are being used. Not a question for the companies, what is the most efficient one to issue, but also taking a step back, these companies are essentially saying this is an existential evolution of our businesses.

7:19We need to invest in this. If I need to pay 10, 20, 30 basis points more than I should have in a normal market, maybe I don't care so much. I'm not that price elastic. So I'm pretty happy to issue more. Maybe investors will feel indigestion because that starts to be too big in terms of single issue limits and things like this. But if you think that that investment is existential for your business, you're going to do it even if the price is a bit high. So to wrap it up, it feels like this issuance is forcing the entire credit market to change and certainly now starting to force the investment grade credit market to change.

7:53Is that how you see it? And what are some things our listeners should be focused on in that realm? So absolutely. The size is getting quite big. Just to give you an idea, in terms of some standard key risk metric that we like to look at in the corporate bond market, duration types, DTS, a QPS creation that many investors are using. The hyperscalers, the six hyperscalers, are now larger than the big six banks, which is incredible because that's happened in less than a year. And so suddenly you have to care about this half a dozen company a lot, as much as you care about the big six. Huge deal for us, generally speaking.

8:25Now, if you look at the impact on the market, it's really what is in the epicenter. That is really what's reacting. What is the closest in terms of risk, in terms of the correlation? So hyperscalers, the AI ecosystem, the tech sector, all this is getting impacted. we are underway to tech sector because of that pressure. We haven't seen the contagion outside of this, but in our perspective, if we continue at that pace, we should start to see an impact on the broader index because you start to have a simple supply demand perspective. There's just too much of that coming in and that's going to make it harder for others to issue just in terms of pure capacity.

8:59So all of a sudden, we're probably going to see more interrelated moves between credit markets and equity markets in the tech space for the foreseeable future as these companies become much more capital heavy. Dominique, it's been great to have you here. This was very eye-opening for me, and I'm sure our listeners are going to really appreciate it, and it's the right time for them to be focused on it. Thank you. To sum it all up, AI-related issuance is quickly changing the complexion of the entire investment grade credit market. This will have impact on the costs of the AI infrastructure build-out and matter for market participants in all asset classes.

9:35To be notified when we release episodes of the Barclays Brief, please hit subscribe wherever you listen to your podcasts.