The AI Debt Boom – How Hyperscalers are Reshaping Global Credit Markets

Intro: 00:00
From Foyston, Gordon and Payne, this is the Foyston for Thought Podcast, bringing context and perspective to the issues shaping markets and portfolios.

Gabriel Lopezpineda: 00:08
Welcome to episode five of the Foyston for Thought Podcast, recorded on July 27, 2026. I’m Gabriel Lopezpineda, Head of Institutional at FGP Investments. In today’s episode, we will be talking about global credit market and specifically how the AI boom is reshaping that market. To discuss the topic, Cam Greenwood, Portfolio Manager of Fixed Income, is here with me. Thanks, Gabe. Good to be here. So when most people think about official intelligence, they think about NVIDIA, ChatGPT, or the next big IPO that may hit the equity market and how that can reshape the AI race. But bond investors see something different. One of the largest corporate financing cycles in decades. Can you explain to us how AI is impacting credit markets?

Cameron Greenwood: 00:57
Sure, Gabe. Yeah, there are really three areas we can focus in on for that answer. You know, one, how the global bond market is, I’d say somewhat quietly, becoming one of the principal funding mechanisms of the AI build out. Two, how and why Canada has become an important part of the story. And three, why some of the most sought-after bond issues of the past year might not have been the most compelling investments, even though everyone realistically bought them. So let’s start with the first one. What does the bond market have to do with the AI build-out? As you said in your question, Gabe, we are currently living through one of the largest corporate financing cycles ever. And it’s also one that most people really aren’t watching because it’s happening primarily in the bond market, not in the stock market. You know, these so-called hyperscalers, you know, think Amazon, Microsoft, Alphabet, Meta, and Oracle, to name a few. Well, they need they need capital and they need lots and lots of it. Now, some people listening might be thinking, you know, but these companies generate enormous amounts of cash flow and hold attractive cash balances on their, you know, really massive balance sheets. And even though that’s true today, almost the unprecedented scale of this investment needed for tomorrow is just so big that they have no choice but to borrow from deep pools of liquidity, i.e., you know, the global bond markets. You know, you know, AI requires this colossal investment in, you know, data centers, fiber, networking, power, cooling, and advanced chips. And it needs it all right now to make the whole thing work. You know, even for these cash-rich companies, the need to issue longer-term debt really is an efficient way to preserve some of their corporate liquidity while diversifying funding sources and to a certain degree, hopefully matching longer duration assets with long-term capital. You know, just to provide some context, Gabe, of the scale required, you know, the hyperscalers, that group of companies I just mentioned, you know, they’re estimated to spend between 800 billion and 1 trillion, that’s with a T in capital expenditures in just 2027 alone. You know, just as a point of reference, the Canadian economy is a little over 3 trillion. So just a massive, massive number.

Gabriel Lopezpineda: 03:18
Now that you’ve mentioned Canada, let’s move on to your second point. What does Canada have to do with the AI story?

Cameron Greenwood: 03:25
Well, Gabe, say hello to the Mighty Maple Bond. You know, something I know you’re aware of, but I’m gonna give a little background for any listeners today that might not be as well versed. You know, a maple bond is simply a Canadian dollar bond issued by a foreign borrower. Think of it this way you know, RBC issues a bond in Canadian dollars, it’s called a Canadian corporate bond. Canadian company, Canadian dollar bond. But if Amazon or Alphabet issues that same bond in Canadian dollars, it’s called a maple bond. Foreign company, Canadian dollar bond. Even though maple bonds have been issued in the Canadian fixed income market for many years now, it was a relatively small or niche corner of the market, because most foreign issuers preferred the deeper US dollar or European credit markets. Also, you know, Canadian investors, we really had limited reasons to prioritize these deals over existing options in our home market. That changed on January 1st, 2025, when newly issued maple bonds became eligible to be added to the FTSE Canada Universe Bond Index. That’s the main benchmark that many Canadian bond portfolios are measured against. You know, that rule change made a real difference. Suddenly, benchmark aware pension plans, insurance companies, ETFs, and mutual funds could now own these securities without introducing tracking error into their benchmark. Meaning, really in plain terms, Gabe, that index tracking firms effectively needed to hold these regardless of price, since the bonds were now part of the index they were built to replicate. With a larger buyer pool, Canada became a much more attractive funding market for global issuers almost overnight.

Gabriel Lopezpineda: 05:06
So that’s maple bonds, but let’s jump back to AI and what’s taken place in the last couple of months.

Cameron Greenwood: 05:12
Yeah, sure. During the first couple of trading days of May, Alphabet launched its first Canadian dollar bond issue, raising a total of $8.5 billion across four maturities extending out to 2056. You know, all told that was the largest corporate bond issue ever completed in Canada. Well, that record lasted all of about one month because on June 8th, Amazon priced an even larger $14 billion five-part maple bond transaction. Once again, maturities extended out 30 years, and in the process, established yet another new record for the Canadian corporate bond market. You know, together these issues really transformed perceptions of the Canadian credit market and demonstrated that Canada could absorb some of the largest corporate bond transactions in the world. You know, Gabe, I’d also note that these deals were only part of a much broader trend across developed markets. The hyperscalers have tapped investors in US dollars, euros, sterling, Swiss francs, and, you know, of course, as we’re just talking about Canadian dollars, all to help finance AI infrastructure. You know, this is really what I meant when I said earlier the global bond markets become one of the principal funding mechanisms behind the AI build-out.

Gabriel Lopezpineda: 06:20
Okay, all that makes sense. But let’s get to your third point. The potential that these bonds weren’t the most compelling investment, because I think that might be quite interesting for our listeners.

Cameron Greenwood: 06:31
Yeah, this last part, it’s critical. You know, the investment decisions around these bond issues. To start, as most listeners will likely be aware, you know, buying a corporate bond is fundamentally different from equity investing. As bond investors, we don’t buy pieces of companies looking to compound value. We manage risk, and we do so through things like careful security selection, understanding capital structures, and you know, trying to avoid deteriorating credits. But we don’t just manage risk, we price it, and we do so using credit spreads. You know, as a result, a great company does not automatically make a great bond investment. When Amazon and Alphabet came to the Canadian market, demand was really extraordinary. Order books were heavily oversubscribed, and the buyer’s list, which is the number of institutional accounts participating in the new issue, that was incredibly long.

Gabriel Lopezpineda: 07:26
And that’s a good thing, right?

Cameron Greenwood: 07:28
Well, yes and no. From a credit perspective, it can be viewed as a strong degree of confidence. But from a valuation perspective, it does raise an important question. Were investors actually being paid enough? Well, the credit spreads on these bonds were wider than government of Canada bonds, you know, as they should be. They were only modestly wider than some of Canada’s highest quality provincial issuers. In certain maturities, they also offered relatively little additional compensation compared with strong municipal borrowers, public sector agencies, and other highly rated Canadian corporate issuers. That doesn’t mean Amazon and Alphabet are poor credits. Far from it. I said this earlier, but both companies possess strong balance sheets, exceptional cash flow generation, and their business franchises are among the strongest in the world. But even the best companies can become expensive investments when investor enthusiasm compresses those credit spreads too far. Gabe, I should probably mention at this point, here at FGP, we didn’t participate in either deal. We believe we were, we just weren’t being compensated enough. Now, we should also probably consider this development from the angle of portfolio construction and active management as well. As I mentioned just a moment ago, virtually every major institutional investor in Canada bought these deals. Active managers, passive managers, pension plans, insurance companies, banks, ETFs, you name it, they bought it. For many, not owning these issues simply wasn’t realistic because they’re benchmark eligible and they are highly liquid.

Gabriel Lopezpineda: 09:07
Well, that creates some interesting questions. For example, if every institutional portfolio owns the same securities, how can you outperform the benchmark in peers?

Cameron Greenwood: 09:16
The honest answer is you probably can’t. Not through those specific holdings, anyways. Active management, it’s never been about finding the most recognizable company or the most exciting story. Instead, it’s about identifying those opportunities where the markets missed price risk. And then we make those disciplined buy and sell decisions in those certain circumstances. Sometimes that does mean owning an Amazon or Alphabet. Sometimes it means allocating capital more towards a provincial issue or a municipal borough or another corporate that offers better compensation for the comparable level of risk. In financing the AI build-out, there’s this really fascinating paradox being created for fixed income investors. It is producing some of the highest quality corporate bonds in the market, you know, issued by some of those the most dominant companies in the world. And yet, because that current demand is just so strong, they’re coming with some of the thinnest compensations for the risk. Passive portfolios, they’re going to keep absorbing it because that’s really what tracking a benchmark requires. But for active investors, you know, the real opportunity lies in looking past those first headlines and asking, you know, different questions. Not so much, is this a great company, but more, is this the best relative value available today? You know, for us, those are two very different types of questions.

Gabriel Lopezpineda: 10:52
This is great insight, Cam, from global bond markets quietly becoming the financiers of the AI build-out, how Canada is playing an important role in that story, and how some of these hot issues may not be the most compelling investments. Thank you for joining us today, Cam. And to everyone listening, thank you. We’ll be back soon with another conversation on markets and investing.

Intro: 11:14
This podcast is intended for informational purposes only and does not constitute legal, tax, security, or investment advice, an opinion regarding the suitability of any investment, nor solicitation of any type. The opinions expressed are as of the date the podcast was recorded and are subject to change without notice. Foytson, Gordon and Payne is registered as a portfolio manager in every jurisdiction in Canada, an exempt market dealer in every province in Canada, an investment fund manager in Ontario, Quebec, and Labrador, and as an investment advisor in the United States. Foyston, Gordon and Payne manages a number of pooled funds referred to as FGP pooled funds that are offered through a prospective exemption to residents of Canada. The values of the FGP pooled funds change frequently. All investment involves risk. Unless otherwise stated, performance is on an annualized basis in Canadian dollars and its gross and management fees. Past performance is not indicative of future performance. This podcast may contain forward looking statements based on reviews, information, and assumptions as of the date of the recording. Listeners are cautioned that actual events may differ significantly. For further information on Foyston, Gordon and Payne, please visit our website at www.foyston.com.