This is the hardest thing I have ever seen - Howard Marks on the AI Frenzy

Alessandro Baroni (0:00)

Every boom arrives with a story, and the story is always good enough to justify the price. The railroads had one, the nifty 50 had one, the internet had one, and now artificial intelligence has one. And here's the awkward part. It might actually be true. That's the bit most people skip past. The story can be completely right, and the investment can still be completely wrong. Both of those things have happened before, in the same market, at the same time. Welcome to Sweat Your Assets. In this episode, learn from Howard Marks, co-chairman of Oaktree Capital Management, and one of the most careful thinkers on risk in this business. For decades, he has been writing memos to his clients about market cycles. Memos that travel a very long way beyond his clients.

And here, he takes on the question everyone is quietly circling. Are markets excessively exuberant? Or is this simply what optimism looks like?

Before it turns out to be right, Marks makes a distinction that most commentary misses entirely.

A bubble, he says, is not a price level. It's a state of mind. It's what happens when investors stop asking whether something is worth what it costs and start believing that no price is too high. Think about that for a moment. It means you can't spot a bubble on a chart. You spot it in the sentences people are using. Now here's the interesting part.

If a technology is genuinely transformative, and this one may well be, does that make the shares a good investment? Not necessarily. A great company and a good investment are two different questions. We collapse them into one all the time. So what does the historical record actually say? We look at where today's valuations sit against it, and at what buying in at elevated multiples has meant for returns over the following decade. The answer is remarkably consistent, even in the stretches where the underlying story turned out to be sound. High prices don't predict crashes. They predict lower future returns, less dramatic and far more useful, which brings us to the practical part. Let's take a step back.

What do you actually do here?

The tempting answer is to go to cash and wait for a signal. But nobody rings a bell at the top. And if someone does ring one, it's worth checking whether they're also charging you for the bell.

Marks' answer is about posture rather than prediction. Where do you sit on the dial between aggressive and defensive? What are you being paid to take risk right now?

Calibration, not forecasting. So if you've been wondering whether to trim, whether to add, or simply whether to hold your nerve, this is the conversation to sit with. Let's get into it.

Howard Marks (2:52)

There's no question about the fact, to use Alan Greenspan's saying from about 30 years ago, that we have exuberance. That's the only thing we know for sure. The question is, is today's exuberance irrational? Number one, I don't think anybody can definitively say so. I've never heard anybody tell me exactly what AI will be able to do, or when, or for whom, or how much profit it will produce, and for whom. There's an arms race going on. I would describe what we call the hyperscalers as mostly the greatest companies I've ever seen. And they're engaged in an arm race. Can only one win? Is it winner take all? Can several win? Nobody can tell me these things. So, I don't think there's an analytical, or what we call a value-based way to decide whether or not to participate in these IPOs and if so, at what price. AI is a concept. We can't define its parameters. And it's a great concept. It's likely to be the most powerful force any of us have ever seen. But that's, I think, that's all we know. And a decision to participate, or not participate, and what price to participate at is, it's really, well, it's what my South African friends call a thumbsuck. You can't put numbers on a pad and figure out what these things are worth, which is what value investors like me historically have done. We have a technological innovation. I've seen several, I've read about many more over the last, let's say, 150 years. This may be the greatest. This may be the most powerful. It's also in many ways the least specifiable. Let's say the railroads back in the 1860s. And then radio in the 1920s, the automobile, computers in the 1950s and 60s, internet in 2000

It may be revisionist history, but I think we had a much better view of what all of those could do.

They didn't have this unimaginable, unlimitable upside that AI has or the, in my opinion, degree of uncertainty. All the things I mentioned were accompanied by what we call bulls. People got excited about developments which were unprecedented. They threw vast amounts of money about building the infrastructure for it.