Let's kick things off this hour with Oaktree Capital's co-founder and co-chairman, Howard Marks. Howard, it is great to have you here.
Kelly, thank you for making the time. Always a pleasure.
You see the big, the big scary graphic. What do you make of the market sell-off today?
You know, one of my favorite memos of my own was written in 2015. It's called, "What does the market know?" And you know, people in my business, we're here to take advantage when we know more than the market. So we shouldn't take our signals from the market. The market exaggerates everything.
So you think that the market doesn't know anything that smart participants don't right now?
Well, I think that's essentially true. I must say I find the AI daunting because it is, in my opinion, the least knowable, least predictable thing I've ever witnessed in my career. I don't think, I don't know how you can figure out, given any one company, how their product will be used five years from now. How many of it they'll sell, at what price, at what cost? What the profit margin will be, which companies will win and lose. You know, I mean, we have unusual uncertainty. I've never seen anything like it.
And others might say unusual potential upside. Yes. When you hear, so you guys have to commit capital every day backing companies across all parts of the economy. Where do you feel comfortable that there will be the ability to repay, and where would you step back and say, for instance, with software? I don't know that, we're not so sure.
There are great companies, there are lesser companies. I think the important thing about AI is, and I said this in a memo in December, if you want to invest in AI, you should probably buy the stocks. At least with the stocks, if things go right, you get a lot of upside.
Do you mean the stocks, for instance, Anthropic and OpenAI aren't public yet. So, do you mean stocks of those people who are making AI models, or just people in the value chain?
You know, I don't know enough about it to get specific as to which companies, but the issue is that when you buy debt, you have no upside. The best thing that can happen is you get the interest and principal repayment that you bargain for. After that, it's all different degrees of downside. So, you should only make loans if you're highly confident you'll be repaid, and if the yield compensates for the uncertainty. But if you're going into a very uncertain situation where you can't say that you're highly likely to get paid, don't lend the money. If you want to play, buy the stocks. At least then you have the upside.
But where, I mean, I think of your firm as the opposite. I think of your firm as being a credit firm.
We are a credit firm. So we're only playing in special situations, mostly relating to AI infrastructure, things like data centers and AI factories.
Are you, are you still backing AI data centers? Do you back neo-cloud business models?
That's different. Neo-clouds are tenants. We're backing long-term projects and the creation of the data centers. Yes. And we're trying to do it in places which are less crowded, where there are inhibitions to the construction of too many data centers. You know, we're now part of Brookfield, and Brookfield has a great niche in power, and I think power is going to be a real bottleneck. When you know they want to increase the data infrastructure three to ten times in the coming years, where are they going to get three to ten times as much power? So I think that's an example. Everything you want to do in the investment business, there are ways to do it safer and less safe. The less safe ones are usually the ones that offer maximum upside. The safer ones, because they're safe, don't have to offer so much upside to attract capital. But that's a fundamental choice. And with something as uncertain as AI, despite its great potential, I'm biased to playing it the safer ways.
But you are playing it, which is to say you don't think the whole thing is necessarily a bubble or an area that is so risky with the build-out that your firm needs to steer entirely clear.
First of all, when people say to me, is it a bubble? I say to them, what do you have in mind? Is AI an illusion? No. AI is a real thing that's going to make a real change in our lives, for better and worse. Is the amount of money that the companies are investing a bubble? I don't know because I don't know what their revenues will be and what rate of return that implies. Are the investments that individuals can make a bubble? Are they overpriced? Is there too much eagerness? Again, I don't know. But you know, until recently, I mean, when you see market wobbles like today, what it tells you, in my opinion, is that there was a lot of optimism, some of which just got deflated.
What about the size with which these tech companies are coming to the market for debt? Has that created any problems for investment grade or high yield credit markets? Has it widened spreads? Do you think it will widen spreads, or is this, as I think Rick Reeder put it the other day, the opportunity of a lifetime? Because if you're a credit investor, you have more to pick from than ever with some companies that many would say seem unlikely to ever default. But that said, you have the example of Meta, which has at various points 70% of revenues for this build-out. So the amount of debt coming into the credit markets from the tech companies, how are you thinking about that?
Well, I think it does give you a lot of choices. I think that the most obviously safe payers aren't offering that much return, you know. So they still involve some technology risk, and I'm not sure that their credits are a giveaway.
