Markets Edition: How Does Debt Demand Affect AI Investors?

>> Research at City Markets Edition. >> Hi, I'm Scott Kronert, Head of US Equity Strategy at City Research. Welcome to Research at City Markets Edition, covering various topics that work within the US Equity Markets and in this session, the Fix Income Markets as well. With me today is Dan Soar, the Head of US Investment Grade Credit Strategy here at City Research. I've invited Dan here today to talk about AI in terms of a bomb investors perspective. Welcome Dan and thanks for joining us. >> Nice to be here Scott. >> All right, so big picture. AI obviously kicking in as an ongoing theme within the markets. From an equity perspective, let's see, the way this has been unfolding is that we've talked about the AI influence part of the S&P 500, approaching as much as 55% of the index capitalization. Now we know one of the big ongoing story lines has been the surge in AI-related catbacks. Generally speaking, we're looking at roughly a trillion by the global AI influence this year. Perhaps tripling if not quadrupling by the end of the decade. So we're looking at some big numbers. Now we know that going into this year, the view was at hey, look at the companies that are spending a lot of this capital are doing it out of free cash, which is great, but as the years unfolded, we've seen the metrics such that free cash generation is now falling significantly, and the pivot has been to debt financing is a way to finance a lot of this cap expending. What this is doing in our view is creating some stress points within the markets that we need to be attentive to, presumably from your perspective, as we continue to look at this AI-related cohort and its influence on the broader markets. So Dan, help us here kick it off of some context around the aggregate size of the AI-related debt issuance thus far and how that fits into your investment grade narrative. Sure, Scott, first of all, the investment grade market in dollars is extremely deep. We issue more than a trillion dollars of debt every year and the market as a whole as more than $10 trillion outstanding. The market trades anywhere in the order of $50 billion a day. And this really creates what I would call the most liquid, spread product market globally. And so it's a market that has the capacity to absorb an enormous amount of debt when there's an opportunity for an industry to build new income generating assets. That's really what's gotten the market very, very focused this year. And what we're seeing is some strains, but not really cracks at this stage. To sort of give you a sense of it, when we look kind of at the on balance sheet funding of the hyperscalers, we've had more than $200 billion issued year to day just from a handful of names. That's running it more than twice the pace of last year for those names. And they're spreading it not only in US dollar markets across tenors, but also in euros and sterling and Swiss Frank, in CAD, in Yen, and now in Aussie dollars. And so it's a story with a handful of names, triggering a massive debt raise across the globe. What's really also quite interesting is the way in which the AI boom is being funded not only in traditional on balance sheet debt, but in data center project financing across both IG and high yield, where we see on the order of $100 billion of financing already. That has begun to strain valuations a bit in the investment grade market, where the index might trade around 80 basis points over treasuries. You know, a 10 or 20 basis point under performance is notable. From an equity perspective, and probably from the issue or perspective, it's a small price to pay for what looks like an incredible opportunity from their eyes. There are going to be ups and downs. There's a lot of variety in the names. There's a lot of different types of risk. There's geographic risk. There's structure risk. There's questions around liquidity, and of course, just regular cyclicality of the economy. And that leaves aside the broader questions around how soon and to what extent, to what magnitude the AI-driven investments are going to create free cash flow growth. So these are some of the questions that I think the market's grappling with right now. There are certainly a lot of big numbers, but we're managing it okay with a bit of strain. Okay. So when you think about this, one of the issues that's coming up for equity investors is this notion of circular financing, where the cap expenders are actually, in many ways, become part of the conviction that there's an ultimate payment structure on the interest expense to kind of keep all of this together. How would you say, at this point, fixed income investors are looking at considering and assessing the circular financing component that is so topical right now? Sure. There are really a lot of questions about how the money's going to get raised, because there's so many different types of assets that need to get built, and there's a lot of dependency. So the dependencies really start with the fact that in order to create a data center, you need power, you need chips, you need land and permits, you also need customers, and one of the things that's really challenging is that, or many of the adopters of AI, the costs associated with developing new products are also substantial, and so the inference costs are also high. There's also kind of a knowledge distribution as well, where firms are attempting to adopt AI into their workflows, develop new workflows, develop new startups, and new approaches, and that also takes time and money. So there's a lot of money that's going to need to be raised, and we are getting a variety of issuance patterns. I think the data center financing is probably the most interesting in terms of the way in which these deals are being structured, because the conventional toolkit is probably getting stretched, as I mentioned, off-balance sheet data center financings are creating a little bit of confusion, and because there's risks here that the IG credit markets are not typically prepared or focused on. So there's tenants, there's guarantors, there's different types of lease structures, there's different amortization schedules, and there's different construction risk. And pricing these deals has created some challenges, and there's definitely been ups and downs, as the market tries to figure out which players are going to have the strongest projects, and which projects are the most secure. But I would say more broadly speaking, the circular financing is something that really comes down to whether the chips can get produced, the data centers can get built, and the cash flow can start to get generated. And I think we're probably at the period of greatest opacity in how this all is shaking out. Over the next few quarters, I think we'll start to get a much clearer sense of the cash flow potential, the cash flow growth rates, and start to see whether all of this will work. So I think the circularity is somewhat inherent, and that we're trying to build something from scratch, and there's a lot of interdependent parts. Okay, so to kind of then cut to the chase on this from my porch anyway. So I guess there's some question ultimately, in terms of how deep the market is for AI data center-related debt, you addressed some of that earlier, that this is still small within the broader investment grade construct out there. But I guess it does beg to question from an asset allocation perspective. You know, I have to contend with the AI influence on US equities, but increasingly when we look at whether it's a 60/40 portfolio or other stock versus bond decision points, increasingly it seems like what we're going to see is the AI influence percolating further and further across the fixed income spectrum as well. Potentially changing the way we think about traditional asset allocation approaches, yet any wrap-up views on that, Dan? Sure, there are certainly a lot of questions around the concentration of risk in AI in a combined stock and bond portfolio. When credit is really seeing a lot of issuance in portfolios that have fixed income exposure, whether it's an IG or a high yield or an insecurities that are seeing a greater concentration in tech. But I think it's also helpful to look at the extent to which tech is one of many sectors and is actually not one of the largest sectors in the market. When we look at the Bloomberg Barkley's global ag corporate index, banks represent 22% of that index in dollars. Insurance represents 7% electrical utilities represent around 9% and consumer non-cyclical represents about 14%. Tech is only about 9% or so. And so the base from which this debt is growing is relatively low. This is not a sector that already dominates the market. And that's really important to highlight because when we have periods of time when the banking sector is issuing a lot of debt all at once and it's already one of the largest in terms of the amount of debt in portfolios, that can create some challenges as investors have to determine what the opportunity cost is of participating in a new issuance when they view that another deal is going to come just after potentially cheaper. Now what's really interesting, Scott, is that if you look at the portfolios of non-US investors in the credit markets, technology is much lower as a proportion of their holdings today than it is for US investors. So in the non-US dollar portion of that global corporate ag index, technology represents only around 3.6% of their holdings. What that means is that probably some of the strongest interest and strongest appetite for exposure to tech and to US tech in particular is coming out of Europe, it's coming out of Asia, Australia, and elsewhere, where there are not really large or existing credit portfolios are not heavily exposed to tech. I would say the global economy is tied very closely to the fate of these AI investments. In other words, if these AI investments pan out well, we could see significant productivity improvements. If there's over-building, there could be some challenges there and that also raises questions around contagion risk and financial stability the longer this goes on. I think maybe that the most interesting perspective that I hear from clients is not so much around the risk of growing exposure to tech companies, but really about the risk to incumbents of a transformative technology like Generative AI and the agentic buildout that we're seeing that's likely to disrupt industries from health. healthcare, to the energy sector, to retail and consumer, to media. This is a technology that has captured so much interest in sea suites in virtually every industry because of its potential to change the way these businesses operate and could potentially leave a lot of downside or incumbents that have built businesses and built competitive advantages that AI could potentially chip away at. So I think investors have to think of this really from both ends, which is one, how much exposure do they want to add in their credit portfolio when their equity book is already increasingly tilting towards tech. But alternatively, can you afford an accredited book to avoid this sector when incumbents in the IG market across other industries may end up needing to spend an awful lot of money on R&D to build and maintain kind of parallel infrastructure while they determine how they can incorporate AI into their workflows or their product lines and potentially see margin compression as pricing pressure drops or pricing pressure increases. Gotcha. Okay. That's awesome. So we're at our time. I just want to wrap this up real quickly. What I'm hearing from you, Dan, is that, yeah, we're going to be contending with AI for time to come now and the equity side and fix income markets. What I think I'm hearing is that there's still sufficient demand out there, but it's going to be, the burden of proof is going to continue to be on the issuer to demonstrate that the cash generation is there to kind of fund the debt issuance and that will be an ongoing decision point. This notion of disruption in terms of what you get out of AI may be at some other part of the market's negative dynamic is something that we're going to have to keep an eye on here, but I'll told it sounds like we're in a place right now where the demand construct isn't sufficiently good place that we can keep an eye on debt and the debt financing for AI, but not lose side of it that it's going to be an ongoing evolving, changing narrative. So thank you, Dan. I really appreciate your time. This podcast was recorded on August 31st, 2026. Thanks for joining us today and be sure to be in the lookout for our next markets podcast featuring Derek Willer city's head of global asset location. Also be sure to watch for our other research at city podcast series, which you can also view on the same channel. Thanks and have a great day. This podcast contains thematic content and is not intended to be investment research, nor does it constitute financial, economic, legal, tax or accounting advice. This podcast is provided for information purposes only and is not constituted in offer or solicitation to purchase or sell any financial instruments. The contents of this podcast are not based on your individual circumstances and should not be relied upon as an assessment of suitability for you of a particular product, security or transaction. The information in this podcast is based on generally available information and although obtained from sources believed by city to be reliable, its accuracy and completeness are not guaranteed. As performance is not a guarantee or indication of future results, this podcast may not be copied or distributed in whole or in part without the express written consent of city. Copyrights 2026 City Group Global Markets Inc, member SIPC, all rights reserved. City and city and arc design are trademarks and service marks of city group Inc or its affiliates and are used and registered throughout the world.