Hello and welcome to The Download. I'm your host, Dave Richardson, and it is Stu’s Day. Once again, not on a Tuesday, but a Stu’s Day nevertheless. As we've established after several years of doing this, Stu, every day is Stu’s Day.

That's right, Dave. Every day is Stu’s Day. I say that at home, although it doesn't get quite the same notice from my family members always. But every day is Stu’s Day.

Really? So, your wife and kids, it doesn't seem to work that well when you use it at home?

At times, very much in line with Stu’s Day. But not every day. Every day is not quite Stu’s Day.

Now, I can tell you, there are no Dave Days at home. Between my wife, my two daughters, and four dogs, I'm at the bottom of the food chain. But I'm glad it works for you at least occasionally. And what was Stu’s Day though, while we're doing it a day late here, you were Stu’s daying around with a lot of your smart friends and colleagues in some of the discussions you have. I guess you have these meetings on about a quarterly basis, and you get together and just talk to each other about what you think's going on in your area of expertise, and you walk away with a big picture view of what's going on around the world. And it's always a good chance to check in with you and see what's top of your mind from those discussions. Because it's always interesting.

Yeah, for sure. We call it our investment strategy committee. I think inside the firm there's 20 or so asset classes that are represented in one form or another. So it's a great time to get together and hear what's going on in each class and where some of the issues that are facing each asset class, how we're thinking about them. Generally speaking, it breaks into stocks versus bonds. And what we tend to do is we start off with our economist, Eric Lascelles, who I know you have on lots of podcasts. But, probably the interesting thing there is that economic activity is pretty reasonable. And I think importantly, when you look out over a 6-to-12-month period of time, the one thing I always find interesting is the odds of a recession in the forecast period. And at this juncture, the odds of a recession are quite low. The second thing which we can get into as we discuss some of the different asset classes is the biggest risk if you have a very high valuation and you can't afford a recession as well. So there's the interplay between what we think fundamentally and what we think's priced into markets. But we always start with Eric to give us the lowdown on where economies sit in different parts of the world. And then we roll into fixed income, which is a lot of discussion that flows from the economy about where inflation is headed and where real interest rates are. Not dissimilar to what we've talked about in the past. The real interest rate is back where it's average. It's at a reasonably attractive rate. So interest rates are in this range, and they've been more volatile, but they've been in a range. And actually along those lines, our technical analyst had some studies on the importance of trend. And if an asset is trending, does it predict the future return potential? On the equity side, when you have a trending asset, it tends to be a predictor of future returns. But it's not quite the same case in fixed income. They tend to be more mean-reverting, more range-bound. So supportive of this idea that interest rates are in this range with a reasonable coupon. And on the headline government bond side, the ongoing belief that should the economy slow, we will get some performance back from our bonds in addition to coupon. That was all quite consistent. Then we discuss spreads, and we talk about EM bonds and we talk about high yield and investment grade. Spreads are quite low. Quite tight relative to history. But again, without forecasting economic turbulence likely to remain that way, we did spend some time talking about the hyperscalers— the likes of Amazon and Google and Microsoft— because they've been issuing a lot of capital, they've been issuing a lot of bonds. And that has pushed their spreads out a little bit, a bit wider than they've been. And they've accessed that capital in the United States. They've come to Canada and what they call a Maple bond, which is when a non-domestic company issues a bond into the Canadian marketplace. Those spreads are pretty attractive relative to where they've been.

Stu, maybe just before we go a whole lot further, the idea of a spread. We started talking about government bonds, which tends to be your base to determine that spread. And then maybe go from there and just explain the basics behind the spread because that'll help connect it to the idea that when you move out to higher risk levels in credit, the reason why the spreads are so tight is because if you don't have a whole lot of fear of recession, then you can be pretty comfortable being out there. And it has to do with the spreads relative to government. So why don't you do a little brief explanation of that?

Yeah, great point. If the 10-year bond in the United States is 4.5%, one way or the other, you're getting your money back from the government. It might involve down the road them having to print some more dollars and that has some implications. But one way or the other, that's the— quote, unquote— risk-free rate. So then companies have slightly more risk on top of that. So if you're Berkshire Hathaway and you're a AAA credit, your borrowing is quite tight in spread, meaning you don't pay very much more than the US government. And then as you go from AAA to AA to single A to BBB and so forth, the spread or the interest rate above the 10-year bond starts to widen. It starts to increase. The traditional hyperscalers have been very good credits because they've been so cash generative. So their spreads have widened not so much because people view them as maybe riskier, but there's been a lot of their bonds for sale because they're funding this big CapEx buildout of data centers. So those spreads have widened, and they could bounce around a little bit while they're issuing more bonds but look a little bit more attractive. And then that was a segue into the equity people who are covering the same stocks discussing the capital investment that's going on on the data center side that's drawing so much growth in the broader stock market. So that led into the discussion about how much CapEx could there be in the next couple of years. It might be as much as $2 trillion in 2026 to 2028. How much of that could come from operating cash flow, and how much of that would have to be financed either through equity or through debt? Our rough estimate is probably $400 to $500 billion in the fixed income market. But what we should point out is that these companies do have a lot of cash flow to service debt. So what we're doing is we're saying they can take their debt-to-cash-flow to around 2 times— there's some nuance to that involving operating leases and things like this—but generally speaking, if you saw a business at 2x debt-to-EBITDA, you'd say that's still right down the fairway of investment grade. So then you get into this notion of how much capital will they have and do they have enough to pay the interest? And the answer is yes. And then you get into the returns that they're getting on their data centers where they're hoping for strong growth with north of 30% margins, which will drive good returns there. And then you can just see how this discussion just evolves over the days, one thing leading to another. We talk about the US consumer, we talk about all sorts of things, but just to finish off on the data center was well, what about the open-source models and how does that impact the hyperscalers? There was one chart that showed if you do a query on Claude or OpenAI, per unit that might cost like $4, and the hyperscaler gets maybe 80 cents out of that $4. You might do it on an open-source model and it may only cost 85 cents, but the hyperscaler is still getting 65 cents. So it's not quite as lucrative, but it's still quite reasonable. And then the next leg of that discussion is how much growth will there be in token usage. And this one is kind of all over the map. There's the ongoing growth of intelligence replacement. Then there's robotics. Then there's driverless cars. On one hand, you could sit there and say the growth could be 100-fold, it could be 25-fold, trying to get some ballpark around that. But if you are one of the hyperscalers, even if it's all open source, as long as there's growth, you're still going to make your money. It might turn out differently for the model providers, but for the hyperscalers themselves they should still get their money. And then for the data center, half the money they need is what they call the shell, the structure, the power. Half the money is for the guts, the servers. One lasts longer than the other, so we get into that discussion. And then each new rendition of servers generates about 4 times as many tokens as the old one. So you have these variables to come into play. You have very strong potential growth, but then you're going to have new servers. So you're working on all these calculations. And, we're still at that stage where the growth prospects are strong, and we are getting a better and better handle on how the different variables could play out as we move forward. And that was much of the discussion we got into. Some of the memory stocks. If you recall, we talked about emerging markets, which had been heavily influenced. Those stocks have corrected. So the traditional thesis on emerging markets— which is better consumption, younger demographics, some stronger earnings growth potential— has that reemerged now that we've dealt with some of these memory stocks going down by 30 or 40, 45%? So there was discussion on all sorts of levels about positioning the portfolio going forward. But you're having these discussions around, do I want fixed income over bonds? How much risk do I want to take? Then within fixed income, you're talking about, do I want high yield? Do I want investment grade? Do I want EM? Do I want duration? On the equity side, do I want the S&P 500, the TSX, Japan, Europe, emerging markets? And there'll be more to come as these discussions finish up over the month. But I think what hopefully you get the flavor for is that there was a pretty fulsome discussion on all these measures. And those are just public markets. That doesn't even really involve private markets where we've had some activity in this firm recently. But that's probably a separate discussion and I'm sure you've had some of those portfolio managers on some calls as well.

I have not been in the meetings for quite some time— many, many years— but I still recall the times when I would attend and just sit in the background. Don't worry, I'm not chipping in with any advice. All those decisions are for the smart people. But you're right about that discussion. I think it's such an underrated element of the entire process of investment management, the idea of coming up with different ideas. I guess you would call it a peer-reviewed system where there's a debate and a challenge around different ideas. But you ultimately come up with a strategy, as you've talked about many times on the podcast. It's not a straight line. Okay, we put everything boom right here because we've somebody's won an argument, and this is the direction we're going. We're going to have different scenarios. And we're going to place a little bit here, a little bit there, and then we'll continuously evolve as we figure out exactly the way it's going to play out. But getting a bunch of smart people in a room with a lot of trust and a lot of open discussion leads to better decisions and better results.

Yeah, it's very collegial. Having the comfort to go down all sorts of different rabbit holes, figure things out, present some commentary that might seem wild, but we should talk through the implications for things. It is a real hallmark, I think, to do that in a very open setting. Of course, you're modest, but your common sense and everything that you see and hear is welcome as well when it comes to these discussions.

Yeah. And that's what I was going to say next. You're sitting at the head of the table of these discussions, and you want them to be open and honest. Build that trusting environment again where people can speak up. Say what's on your mind. I don't know how many people who listen to the podcast regularly have walked up to me and said, Stu is so nice when you ask him stupid questions. He's always so patient. So, when people are actually raising good points, you must be incredibly patient and listen very closely when people are raising good ideas.

Well, 100%. And while I much prefer the Stu's day moniker, my nickname in university was Stupide. So, you need to be very open-minded.

Okay. Well, I'll tell you what that means in French if you haven't figured that out. But I think a lot of people are always interested in AI and what's going on there. So the part about the hyperscalers and I just get a sense from what you were saying there. If we talk about what these hyperscalers are getting out of the expansion of AI, I always bring it back to my pie. I got my pie. I'm getting a sliver of the pie. And even if I get a smaller piece, if the pie is gigantic, I'm doing much better. But it seems there's still a question about just how big that pie is going to be. And you're never going to know. We got different scenarios that could play out. We're never going to know exactly how big the pie is going to be, but it seems we're just at a stage right now where there's a lot of questions. Some people think the pie will be the biggest pie in the history of the world, and others think it'll be a really nice-sized pie but not quite as big as maybe some people expect it to be. And that creates the questions around what's going on in AI, and particularly with that build-out.

Yeah, I think that is a very good point. Different than some things as they unravel or as they unfold, the price of a token, that's the cost to do something. And there might be a very significant list of things to do that might be quite dependent on the price of a token, which also makes the demand that much harder to determine. If token prices are X and the hyperscalers make money at that X, then you might say, well, demand will be Y. If it's a half X, it's possible that it's more than 2Y. Because the lower price might then open up more things to do with the tokens, and the hyperscalers would still be good in that situation. So when you think about driving in a Waymo, when you think about humanoid robotics, when you think about the queries that different models might ask that AGI, the general intelligence, that is the hard thing because I might turn on a humanoid robot to do some complex task if tokens are X. I may turn it on to clean my house if tokens are Y, or I might do it myself. So there is a lot to think about in terms of how this will move forward.

Yeah, we know it's going to be a big thing. But as you say, just how much it permeates the world and then how quickly is another factor as well. I should mention the Waymos, for people who haven't been in a city with a Waymo, that's a driverless taxicab. Funny-looking car drives up beside you. It's got a thing on the roof. It's an odd-shaped vehicle. And you look over and there's no one driving it. But there is someone else in there, just typing away or on a call or doing whatever— maybe listening to this podcast, which you can do if you subscribe wherever you listen to podcasts. Or subscribe and watch us on YouTube. Then, you could sit in the cab and someone's looking at you while you're watching us. It would be an interesting whole thing there. But the Waymo or the scope of it is just unbelievable where it could end up. And you say, it likely does end up with the Jetsons where Rosie's cleaning the house in the sky.

Yeah, and I think what we're really thinking about is, do I have a funnel where I'm going to collect token usage. If I'm one of the big companies, that's probably a pretty big funnel. If I'm Dave's data center, that funnel may not be quite as big, right? So then we've got this funnel of demand, and then we've got the data centers, and then there's going to be this significant competition between some of the brand name semiconductors that want to provide tokens, and then some of the companies are getting into designing and developing their own tokens. So you're going to have this funnel of demand, and then you're going to have the cost that you have to provide the token. And the lowest cost token and the biggest funnel is likely going to emerge as the winner in these discussions. And that's just something that we'll be very focused on as the future unfolds.

Yeah, and that's what makes that area of the market so interesting, along with the movements that you see in these stocks, which are quite stunning. We've talked about that quite a bit on previous episodes. The other thing that struck me in your initial comments coming out of the discussions was the idea that if they're borrowing money to do this at a high level or issuing stock to do this, you got to think of it one way versus where it's been up until now where they're just generating enough cash flow to pay for this build-out. And that's really where interest rates come in and are particularly important. And it's been a really unusual fixed-income market. I've described it as dull. Dagmara gets very angry at me when I do that. Bond people think the bond market is very exciting, just like they think a rainy day in Toronto is beautiful weather. They're like that, fixed income people, kind of dour. But the idea that it's been government plus a little bit, in terms of spreads, and that's not a particularly exciting market for bond investors.

Yeah, when I started in the business, I was clipping coupons. What that meant was, you used to get the bond certificate in the mail, and then it would have coupons along the bottom, kind of something you'd see on a telephone pole, like, I lost my dog, call this phone number. So the coupons would be along the bottom of the bond certificate and you'd cut them off and that's how you'd get your interest. In an environment where the 10-year, in the United States anyways, is 4.5%, real interest rates are 2 or 3%, corporate spreads are picking up 1 to 1.5%, the income is quite attractive relative to what it's been historically. And then you have some gyration on days that maybe the economy is a little bit too hot, maybe it's a little bit too cold. So you have that range-bound component we talked about. But if we have any doubts around the economy, there's room for rates to fall. And that's where you would pick up a little bit more cushion from the bond exposure than you might have historically. But it has not been as an exciting place as the equity market for a period of time during rising inflation and strong nominal growth. But it still has that role in the portfolio.

And so you leave these discussions ultimately with an idea around all the different asset classes and what you might do there. Is it a case where you go and you execute those trades immediately? Are you particularly active after that? Or this is something where you're going to implement that strategy as it plays out? What's the discussion to any action that you have? What's the space between those two moments?

Well, this could lead to action, but at the same time, we have an investment policy committee that meets ad hoc all the time. So that too could lead to a change. We have lots of ways to make changes. I think what this does is this is a very good day and a half to set the table for the next 3 months and to go back and revisit that and say how has our thinking evolved over that 3-month period? This could lead to change. But in all honesty, we could meet tomorrow and make a different change. Changes are very fluid. This is really a level setting of thinking from quarter to quarter. We can go back through time, we can feed it to AI. We could say, how did you think about this in the past? We can go revisit notes about what did it look in 2015 when this presented itself, these types of things.

Yeah, and that's the one thing about the market and what makes investing so interesting, especially on a professional level, just how fluid it is day to day, minute to minute. News flow. It shifts things, changes perspectives, changes the way people think about the value of something. Some people think it's going up, some people think it's going down. That makes a market. Buy, sell, and away we go. Stu, we’re obviously not going to share with the entire world precisely every discussion that's happened, but I thought it would be really nice to give a window into the way that a bunch of professionals get together in this business and share ideas and have it come to, as you say, a level set from which you're going to move forward. You're going to make decisions again on a regular basis, but it's that level set opportunity. Then you move on from there with a good idea about what everyone's thinking on everything that's going on in the world and how it might affect your particular asset class or the money that you're managing.

That's right.

Wow, I actually got that right. I was worried. I've been on a dumb question run on this episode of the podcast today, Stu. I think we need to get back to Stu's Days. I'm sharper on the actual Stu's Day.

Yeah, that's right.

You seem skeptical, so maybe not. If Stu's Day is every day, maybe I'm not that sharp on any Stu's Day. Anyway, Stu, thanks a lot. I know how busy you are, and I really appreciate you taking the time this afternoon to get together. Say hi to everybody, and we'll talk to you next week.

Great. Thanks, Dave.