Are AI valuations creating a bubble?

Original Release Date: December 3, 2025

Our Global Head of Fixed Income Research and Public Policy Strategy Michael Zezas and Chief Global Cross-Asset Strategist Serena Tang address themes that are key for markets next year.

Read more insights from Morgan Stanley.

----- Transcript -----

Michael Zezas: Welcome to Thoughts on the Market. I'm Michael Zezas, Global Head of Fixed Income Research and Public Policy Strategy.

Serena Tang: And I'm Serena Tang, Morgan Stanley's Chief Global Cross-Asset Strategist.

Michael Zezas: Today we'll be talking about key investor debates coming out of our year ahead outlook.

It's Wednesday, December 3rd at 10:30am in New York.

So, Serena, it was a couple weeks ago that you led the publication of our cross-asset outlook for 2026. And so, you've been engaging with clients over the past few weeks about our views – where they differ. And it seems there's some common themes, really common questions that come up that represent some important debates within the market.

Is that fair?

Serena Tang: Yeah, that's very fair. And, by the way, I think those important debates, are from investors globally. So, you have investors in Europe, Asia, Australia, North America, all kind of wanting to understand our views on AI, on equity valuations, on the dollar.

Michael Zezas: So, let's start with talking about equity markets a bit. And one of the common questions – and I get it too, even though I don't cover equity markets – is really about how AI is affecting valuations. One of the concerns is that the stock market might be too high, might be overvalued because people have overinvested in anything related to AI. What does the evidence say? How are you addressing that question?

Serena Tang: It is interesting you say that because I think when investors talk about equities being too high, of valuations – AI related valuations being very stretched, it's very much about parallels to that 1990s valuation bubble.

But the way I approach it is like there are some very important differences from that time period, from valuations back then. First of all, I think companies in major equity indices are higher quality than the past. They operate more efficiently. They deliver strong profitability, and in general pretty solid free cash flow.

I think we also need to consider how technology now represents a larger share of the index, which has helped push overall net margins to about 14 percent compared to 8 percent during that 1990s valuation bubble. And you know, when margins are higher, I think paying premium for stocks is more justified.

In other words, I think multiples in the U.S. right now look more reasonable after adjusting for profit margins and changes in index composition. But we also have to consider, and this is something that we stress in our outlook, the policy backdrop is unusually favorable, right? Like you have economists expecting the Fed to continue easing rates into next year. We have the One Big Beautiful Bill Act that could lower corporate taxes, and deregulation is continuing to be a priority in the U.S.

And I think this combination, you know, monetary easing, fiscal stimulus, deregulation. That combination rarely occurs outside of a recession. And I think this creates an environment that supports valuation, which is by the way why we recommend an overweight position in U.S. equities, even if absolute and relative valuation look elevated.

Michael Zezas: Got it. So, if I'm hearing you right, what I think you're saying is that comparisons to some bubbles of the past don't necessarily stack up because profitability is better. There aren't excesses in the system. Monetary policy might be on the path that's more accommodative. And so, when compared against all of that, the valuations actually don't look that bad.

Serena Tang: Exactly.

Michael Zezas: Got it. And sticking with the equity markets, then another common question is – it's related to AI, but it's sort of around this idea that a small set of companies have really been driving most of the growth in the market recently. And it would be better or healthier if the equity market were to perform across a wider set of companies and names, particularly in mid- and small cap companies. Is that something that we see on the horizon?

Serena Tang: Yes. We are expecting U.S. stock earnings to sort of broaden out here and it's one of the reasons why our U.S. equity strategy team has upgraded small caps and now prefer it over large caps. And I think like all of this – it comes from the fact that we are in a new bull market. I think we have a very early cycle earnings recovery here. I mean, as discussed before, the macro environment is supportive. And Fed rate cuts over the next 12 months, growth positive tax and regulatory policies, they don't just support valuations. They also act as a tailwind to earnings.

And I think like on top of that, leaner cost structures, improving earnings revisions, AI driven efficiency gains. They all support a broad-based earnings upturn. and our U.S. equity strategy team do see above consensus 2026 earnings growth at 17 percent. The only other region where we have earnings growth above consensus in 2026 is Japan; for both Europe and the EM we are below, which drive out equal weight and slight underweight position in those two indices respectively.

Michael Zezas: Got it. And so, since we can't seem to get away from talking about AI and how it's influencing markets, the other common question we get here is around debt issuance related to AI.

So, our colleagues put together a report from earlier this year talking about the potential for nearly $3 trillion of AI related CapEx spending over the next few years. And we think about half of that is going to have to be debt financed. That seems to be a lot of debt, a lot of potential bonds that might be issued into the market – which, are credit investors supposed to be concerned about that?

Serena Tang: We really can't get away from AI as a topic. And I think this will continue because AI-related CapEx is a long-term trend, with much of the CapEx still really ahead. And I think this goes to your question. Because this really means that we expect nearly another [$]3 trillion of data center related CapEx from here to 2028. You know, while half of the spend will come from operating cash flows of hyperscalers, it still leaves a financing gap of around [$]1.5 trillion, which needs to be sourced through various credit channels.

Now, part of it will be via private credit, part of it would be via Asset Backed Securities. But some of it would also be via the U.S. investment grade corporate credit bond space. So, add in financing for faster M&A cycle, we forecast around [$]1 trillion in net investment grade bond issuance, you know, up 60 percent from this year.

And I think given this technical backdrop, even though credit fundamentals should stay fine, we have doubled downgraded U.S. investment grade corporate credit to underweight within our cross asset allocation.

Michael Zezas: Okay, so the fundamentals are fine, but it's just a lot of debt to consume over the next year. And so somewhat strangely, you might expect high yield corporate bonds actually do better.

Serena Tang: Yes, because I think a high yield doesn't really see the same headwind from the technical side of things. And on the fundamentals front, our credit team actually has default rates coming down over the next 12 months, which again, I think supports high yield much better than investment grade.

Michael Zezas: So, before we wrap up, moving away from the equity markets, let's talk about foreign exchange. The U.S. dollar spent much of last year weakening, and that's a call that our team was early to – eventually became a consensus call. It was premised on the idea that the U.S. was going to experience growth weakness, that there would also be these questions among investors about the role of the dollar in the world as the U.S. was raising trade barriers. It seemed to work out pretty well.

Going into 2026 though, I think there's some more questions amongst our investors about whether or not that trend could continue. Where do we land?

Serena Tang: I think in the first half of next year that downward pressure on the dollar should still persist. And you know, as you said, we've had a very differentiated view for most of this year, expecting the dollar to weaken in the first half versus G10 currencies. And several things drive this. There is a potential for higher dollar negative risk premium, driven by, I think, near term worries about the U.S. labor markets in the short term. And as investors, I think, debate the likely composition of the FOMC next year. Also, you know, compression in U.S. versus rest of the world. Rate differentials should reduce FX hedging costs, which also adds incentive for hedging activity and dollar selling.

All this means that we see downward pressure on the dollar persisting in the first half of next year with EUR/USD at 123 and USD/JPY at 140 by the end of first half 2026.

Michael Zezas: All right. Well, that's a pretty good survey about what clients care about and what our view is. So, Serena, thanks for taking the time to talk with me today.

Serena Tang: And thank you for inviting me to the show today.

Michael Zezas: And to our audience, thanks for listening. If you enjoy Thoughts on the Market, please leave us a review and share the podcast. We want everyone to listen.