Markets and Economy Above the Noise: Perspective on bubble fears

Are we predisposed to see bubbles everywhere?

I think of my grandparents’ generation who came of age during the Great Depression and became lifelong savers and long-term skeptics of the stock market. Their formative experience shaped how they invested for decades.

What if the same thing has happened to us?

Many investors today may have been shaped by the technology bubble in 2000 and the housing bubble in 2008. Have those experiences conditioned them to see a bubble whenever investment, enthusiasm, and technological change accelerate together?

Artificial intelligence is a good test. Investment in it has been surging,¹ but so have earnings.² Demand for chips and memory has been straining supply.³ AI model providers have been generating growing revenues,⁴ and we’re still early in agentic AI, although swarms of AI agents have been working on our behalf.

Maybe this will one day prove to be a bubble, but to me it’s hard to argue that day is today. The market’s price-to-forward earnings ratio, a valuation metric, has been essentially unchanged from five years ago, and is roughly one third of what it was at the peak of the tech bubble.⁵

A lot of investment isn’t evidence of a bubble. The question is whether that investment has been excessive relative to the opportunity. Is it excessive relative to today’s revenues? Debatable. What about the potential addressable market opportunity, the productivity gains AI could produce, and the larger economy those gains could help create? That seems less likely.

Perhaps a greater risk is being so determined not to repeat 2000 that we fail to recognize something genuinely different.

It may be confirmation bias but…

… the spike in Treasury yields this year⁶ was likely not primarily about concerns over the sustainability of US debt, or the bill finally coming due, as Stanley Druckenmiller, Ray Dalio, and any number of other billionaires warned. If that was the story, wouldn’t we have expected the rise in yields to include not just higher real yields, but also higher inflation expectations?⁷ After all, if markets were genuinely concerned about debt sustainability, inflation would be one potential path out of the problem.

Instead, more recently, increased investor scrutiny of the AI trade, with CoreWeave as a case in point, has coincided with a rally in US Treasuries.⁸ So much for the debt spiral and the supposed end of US Treasuries as the world’s flight-to-quality asset.

Since you asked

Q: Was the spike in borrowing costs for CoreWeave the beginning of the bursting of the AI bubble?

A: No, I don’t think so. CoreWeave was a reminder that not all AI is created equal. Its borrowing costs surged⁹ as investors questioned whether a highly leveraged neocloud could generate sufficient returns on its enormous infrastructure investment. The research firm Gartner, Inc. went further, forecasting that more than 80% of neocloud providers could cease operating independently by 2030.¹⁰ Perhaps, but 2030 is a long way away, particularly given how quickly AI has been evolving.

What mattered to us was what didn’t happen. Credit spreads, in the investment grade and high yield bond markets, didn’t widen significantly.¹¹ It appears that investors weren’t demanding anything resembling CoreWeave’s protection costs from the hyperscalers, whose diversified businesses may leave them better positioned to finance AI investment.¹² And the S&P 500 finished the week ended October 16 just 2.5% from its record high.¹³

That’s hardly a bubble bursting in my view. It’s more of a market that appears to have separated the stronger players from the potentially weaker ones, precisely the kind of discipline I think the AI naysayers should applaud.

It’s that time again — midterm elections

Midterm elections are upon us. Every two years, investors wonder whether it’ll be an election that meaningfully changes the market outlook. History suggests that it rarely is. Economic growth, earnings, inflation, and Federal Reserve (Fed) policy have tended to matter far more than which party controls Congress.

Nor is the most likely outcome likely to be a major surprise. Markets have had plenty of time to contemplate the possibility of divided government. It would make another large fiscal package more difficult, reduce the likelihood that the election would meaningfully change the trajectory of economic growth, inflation, or Fed policy.

History provides some additional perspective. Since 1960, the president’s party has lost an average of 22 House seats in midterm elections.¹⁴ Yet the stock market has gained an average of 16% over the following year, with none of those periods delivering a negative return.¹⁵ That may be more coincidence than anything else, but it certainly doesn’t suggest something investors should fear.

It was said

“I don’t even want to be Steve Eisman.”

– Steve Eisman, investment analyst and portfolio manager

Eisman is best known for calling the housing crisis and profiting from it, a story famously told in Michael Lewis’s The Big Short. It was fitting that he was asked about the cottage industry of investors trying to become the next Steve Eisman by predicting the next great crisis. His response was a well-placed jab at those perpetually searching for the next disaster.

Eisman acknowledged legitimate concerns about the amount of AI infrastructure being planned relative to the revenues these models may ultimately generate but was humble enough to admit that he doesn’t know how it will play out. More importantly, he stressed how little today resembles 2008, particularly given the relative health of the financial system.

I truly believe Eisman has no desire to be the next Steve Eisman. Although, to be fair, it must be great to have Steve Carell play you in a movie.

Phone a friend

We’ve received a lot of questions about the recent rise in short rates and the implications for the credit market. I reached out to Scott Baskind, Invesco’s Head of Global Private Credit. His response:

“Higher rates have increased the income potential of senior loans in my view because of their floating rate structure. But there’s another side to that. Higher borrowing costs have put additional pressure on more highly leveraged companies, particularly those whose capital structures were built when rates were much lower. So, we think this is an environment that could reward selectivity. And that selectivity is increasingly important.

We’ve seen a real divergence between higher quality borrowers and the weaker parts of the market. Stronger companies with durable cash flows and appropriate capital structures have generally held up well, while more leveraged borrowers are feeling the pressure of higher financing costs. The good news is that we don’t feel compelled to reach for that additional risk today. We believe senior loans can generate attractive income potential and maintain a bias toward quality, which could be a compelling combination in this environment.”

Think/Rethink: Where familiar beliefs meet inconvenient evidence

Think: Investing during a Fed tightening cycle is a bad strategy.

Rethink: Historically, stocks have performed better in the years following hikes than cuts because the Fed typically raised rates when nominal growth and corporate profits were strong.¹⁶

On the road again

A recent trip took me to Villanova, Pennsylvania, where an investment professional challenged me on my optimism around AI, pointing out that asset managers have every incentive to promote the story. “Not necessarily,” I said. (I’m sure the Invesco market leader in Villanova would be just as happy if I explained how a high-quality bond portfolio might fit into his clients’ portfolios.)

I told him I think about AI across three horizons. First, the near-term buildout. Next, a period of excess and overvaluation, which I believe would eventually unwind. And finally, the longer-term structural growth story, which I believe could transform most facets of the global economy and our lives. Importantly, AI hasn’t killed the business cycle, and I have little doubt that the excess will eventually emerge. I just don’t think that moment is now. Until then, I’ll stick to my usual knitting and watch the indicators (leverage, inflation expectations, credit spreads, lending standards, etc.) that have historically helped indicate if a cycle may be nearing its end.

Later that evening, he confessed that he wasn’t bearish on AI either. He just wanted to see how I’d respond. Hopefully I passed his test. These are the client interactions that keep the job interesting, even well into my third decade of doing it.