Key Takeaways
- Bank of America’s Sebastian Raedler says markets are mistakenly pricing the AI trade as an ongoing certainty.
- The strategist sees scope for a 5%-10% pullback as his base case, while a deeper recession could see declines of 40%-50%.
- Raedler remains underweight on European equities, but flags opportunities in defensives and domestic cyclicals.
Karen Gilchrist: Stock markets have proven remarkably resilient so far in 2026, buoyed by AI exuberance and expectations for a supportive economic backdrop. But some are warning that investors have become overly optimistic amid mounting risks. To discuss why markets appear excessively bullish, the possibility of a sharp pullback, and what it all means for European equities, I’m joined by Sebastian Raedler, head of European equity strategy at Bank of America.
Sebastian, thank you for joining me. Now, you’ve been saying for some time that markets do appear excessively optimistic, and that these risks are mounting around us. What is your explanation for this, what are you seeing there, and what risks do you think investors are overlooking?
Markets Are Pricing Certainty in the AI Trade—History Suggests Otherwise
Sebastian Raedler: I think what we’re living through—and I think you already hinted at this in your introduction—we’re living through one of the big capital cycles, one of the big investment splurges. You really have to go back over the last 200, 300 years, maybe you find five or six episodes where you’ve got so much capital coming in. And it explains to us, because, as you say, we’ve been cautious for a while, and we’ve continuously been surprised by how the system, which over the last 20 years was so fragile, how it has taken all these shocks—the trade war, the Iran-US war.
Earlier this year, the International Energy Agency told us this is the worst energy supply shock, and we never really saw global macro surprises turning negative. And it’s becoming increasingly clear there’s a wave of money basically washing through the system, as there’s this strategic arms race, and you’ve gone basically from spending an annualized $300 billion—which is maybe 1% of US GDP—on this AI infrastructure, to now the expectation by consensus that this would be $900 billion of spending. And, of course, it buoys profit margins, it helps economic growth, and it just makes the system a lot more resilient.
Ultimately, it all comes down to this one question: is this sustainable? And that translates into the question: will the companies—the hyperscalers who are spending this—make an adequate return on that investment? And I think the answer is it’s plausible that they will. But the market is pricing that as a certainty, while history teaches you that often these booms turn into busts. So right now you’ve got a market that is pricing all the good news—no risk premium, very elevated margin expectations—and I’m saying if you look at a probability-weighted distribution, there are many ways in which this can go wrong. All the good news is already in the price, but the way in which it could go wrong is ignored, basically, at the moment.
Stock Market Drivers Appear Stretched
Karen Gilchrist: Right. Is there a particular area that you’re most concerned about? Would that be excessive valuations, or is it these earnings expectations that maybe you think are not so realistic?
Sebastian Raedler: I would say, in general, there are three drivers for the equity market. There are the earnings expectations. There is the risk premium. And there’s the risk-free discount rate. Two of these three look basically to be in an unusually stretched position. So you’ve got record profit expectations on record margin expectations. I’m saying that with the expectation that, at the global level, 25% of earnings will be contributed by the tech sector. That number on average has been 15% over the past decade. At the peak of the dot-com bubble, it was 13%. So you’re saying all these bets with certainty are going to pay off? And I’m saying, sure, they might pay off, but there’s a very wide probability distribution. And so you’re already saying there’s going to be tremendous profitability here.
Now, we can discuss the challenges to this. So, it’s a very competitive field. Will you make a high margin in a very competitive environment? Normally, that’s not the case. And you have tremendous depreciation charges that are basically because if you invest a lot, right now you get a profit boom, because everybody is getting revenue from the spending. Nobody’s getting the costs because the costs get capitalized; they go through the balance sheet, but they will show up again as depreciation charges in the future. It’s very hard for capital-intensive businesses to have high margins because the depreciation charges push against it. The current expectation is that all of this received financial wisdom is irrelevant; it’s a new magical realm of profitability. And I say maybe, but a lot has to go right.
So that’s the first thing that is stretched. The second thing that is stretched, and this is normally the main driver of equity markets, is risk premia. What are the risks that you want to get compensated for if you’re in financial markets? And some clients are smiling at me, and they’re saying, ‘Look, Sebastian, you have to accept that there are risks, because otherwise you’ll always be on the sidelines and you will not be participating in this.’ And I’m saying, ‘Sure, I’m happy to take risks, but I want to get compensated for these risks.’
So we were, for instance, long European equities in March 2020, when there were a lot of risks, but the compensation was very high. Right now, risk premiums in credit markets and in equity markets are at a 20-year low. This is a market that’s saying there are no more risks out there, and I’m just saying that is maybe too optimistic.
A 5%-10% Stock Market Pullback Is the Base Case
Karen Gilchrist: So in your scenario, or this potential scenario, do you have a sense of the magnitude of a potential downturn? What might we be looking at, over what time frame, and are there any historical parallels you can draw?
Sebastian Raedler: I would say there are two scenarios I would think about. The first is our base case. We would just say there are still risks around, even if they’re not priced. And I said there are these three drivers: profit expectations, risk premia, and the risk-free discount rate. Interestingly, the risk-free discount rate is actually where you can make the last remaining bull argument, because they are at close to 20-year highs. And so if you, let’s say, get a fade in inflation. If you, for instance, look at what the new Fed chair, Kevin Warsh, is waiting for, he’s saying we’re going to get productivity-driven disinflation. And if you get that, central banks maybe can become less hawkish, then real bond yields can come down. That’s really the last remaining driver of equities that is not yet at a massively bullish edge, effectively.
Therefore, I would say we are, of these three drivers, most focused on the risk premium. We say there are still risks out there. We’ve got virtually zero jobs growth in the US. We still have a Strait of Hormuz [blockade] and basically energy supplies that are constrained. We’ve been running down our inventories. That’s not a sustainable situation. I want to get a risk premium to be there. I’m not getting that. And there are big questions of whether AI monetization will happen. The market is pricing that as a certainty. I’m saying there are ways this can go wrong. So, I’m saying in my base case scenario, maybe risk premium, which is currently at a 20-year low, maybe up 100 basis points. That’s a 5%-10% pullback.
You said at the beginning a ‘sharp’ pullback. 10% after a stronger run, is that sharp? Is that just a bump in the road? Some clients are even saying if it’s only 10%, I’m not interested. Call me if you’ve got a more interesting story. But then there’s also the fact that the main driver of equities is the business cycle. And it’s easy to lose sight of that in the excitement. We have a very mature business cycle. We have a very long recovery. The unemployment rate has troughed and started rising. Historically, that has been a major red flag.
Right now, people are acting as though the business cycle just doesn’t matter at all. Everything has been propped up by the growth-enhancing and rejuvenating powers of AI investment. If, as our global strategists are saying, we might get surprise capex cuts from the AI universe, for instance because the bond market is no longer willing to underwrite all of this, then this growth driver booster goes into a dampener. If the business cycle ends, which is not a base case, but we have to think about this because we are in such a mature state, then of course the downside is not 10%. Then you get a proper pullback. That’s not our base case, but we have to be aware that the logical sequence of business cycle dating would say that’s the next chapter that is waiting.
Karen Gilchrist: Have you forecasted what that could be then?
Sebastian Raedler: So I think in a normal recession—and I want to make it very clear that it’s not our base case, it’s just that we have to be aware that at some point when these boosters are running out the probability of that will come up—a typical recession will see a pullback of 40%-50%. And I would say as a macro-aware investor, I will always be far happier to be bullish on the equity market when I say, look, now the margins have gotten compressed, now we’ve got a rebound, now the risk premia are very high, some risks can be priced out. Right now, there’s really none of that potential because all of these things have just happened to such a powerful degree already.
AI Capex, US Labor Market, Energy, and Credit as Key Risks
Karen Gilchrist: If your concerns prove correct, do you have a sense of where the cracks might emerge first? Would it be hyperscalers, data centers, energy infrastructure?
Sebastian Raedler: I think there are basically maybe four areas of vulnerability that I think are underpriced. So clearly there’s the question: Will you be able to monetize? The bulls, especially after Q2 results, will say, ‘We see all this AI revenue.’ The pushback against all this is we still have a lot of VC funding that is sustaining, especially the activity. So if you look at Microsoft MSFT, a large part of their order backlog is from OpenAI. And we know OpenAI at the moment is not a profit-making company. How long will you be able to sustain it? And it’s basically a race. How quickly will this funding run out versus how quickly is the end demand going to scale up? And my concern that I have is OpenAI is now competing in a way that they did not expect against a lot of cheap, very powerful competitors.
There’s a very interesting interview by Sam Altman in early 2025, where he said, ‘Nobody needs to compete with us. We’ve already won this race. We have the best models.’ And I think if you look at the competitive landscape today, you would say that is an overstatement. Will you be able to make money and extract basically a margin in a field that is so competitive, so fragmented, so crowded?
So the first question is: Will you be able to scale up this AI investment, or will there be roadblocks from monetization, credit funding availability, or political pushback? In many states in the US, you get pushback effectively. If you want to scale hyperscaler capex spending to, let’s say, $1.2 trillion next year, you have to convince enough communities to let you do that. If not, you will not get that growth, and you will not be able to distribute there.
Therefore, either we get a disappointment in terms of negative catalysts on this AI investment spending, or we’ve seen surprising renewed weakness in the US labor market. Historically, you have never seen a recovery survive with negative payrolls sustained over a prolonged period. Right now, the margin of safety on that front is effectively zero. And if you get a renewed spike in energy prices, the US consumer is very vulnerable, especially the US consumers who powered through that crisis but are very stretched. The savings ratio is effectively at an all-time low. As a macro person, that makes me very nervous.
Both Pimco and our own credit strategies say there will be more defaults coming, because you’re in a higher-for-longer rate environment, you get pressure on private credit, you’ve got a lot of fragilities in this AI funding. Current credit spreads and risk premia in the credit market are saying zero defaults are priced. If you get defaults, that again would be a repricing. So four areas of vulnerabilities are not priced: the sustainability of the AI capex boom, the US labor market, energy supplies, and the kind of default credit kind of stability situation.
Opportunities Exist in Defensives, Especially Beaten-up Consumer Staples
Karen Gilchrist: Where should investors be positioning themselves in the face of these risks? Where do you see the most discounted opportunities?
Sebastian Raedler: So, you have some of the defensive safe-havens in the market that have gotten utterly crushed. European staples have been a dog of a performer, underperformed more than 30%, really just going down in a straight line. If you look at that chart, you would think that business model is truly broken. And some people are arguing that. They say consumers are turning away from sugar, they turn away from alcohol, the GLP-1. I speak to my analysts, and they’re saying, ‘No, our business is fine. Underlying organic growth is fine. It’s just that the sector performance-wise doesn’t work in this environment, because we get the least benefit from this AI investment boom.’
In fact, on our correlation analysis, staples is the sector with the largest negative correlation to any type of AI proxy. And so, for me, this really looks like a loaded spring. This sector is not structurally broken; it is just cyclically challenged. It will always do poorly in an environment of strong growth and very tight risk premia. And I would simply say the range for credit spreads, typically, is 300-1,000 basis points. Staples always outperform when risk premia widen. Currently we’re at 260, basically at the very bottom of that. Any risks priced into the market will be a fountain of youth for the staples trade.
Karen Gilchrist: We did see some of that rotation when we saw that selloff in AI stocks going into European equities. European earnings growth has improved. But are you positive on European equities versus global? How are you thinking about that?
Sebastian Raedler: So we are still negative. They’ve bounced a little bit. They, of course, have underperformed dramatically over the past decade, mainly because Europe has been an economic underperformer relative to the US. If I say the main concern really is the fragility of the AI trade, that, in principle, should make me more positive on European equities. The problem is twofold. We still think that they’re going to be a structural underperformer, because—there’s now a little bit of a cyclical bounce happening in Europe—but generally we think Europe is still going to be an underperformer. But, more importantly, we think in the environment that we’re describing, there’s scope for bond yields to start fading.
There are effectively two types of bears out there. There are the ones like me who say risk premia are just too low. The others say there’s a lot of upside for rates and inflation, that there’s going to be more hawkishness. In our base case, as some of these supply disruptions from the war start to fade, together with the dollar strength we’ve seen, there’s scope for inflation to fade. If the labor market is more fragile, that’s going to take the wind out of the hawkishness. So we are positioned for lower bond yields.
And the problem with Europe is it’s a value trade; it needs bond yields to rise. And so we expect lower bond yields and structural underperformance to still weigh on Europe for it to continue that recent underperformance. It of course makes me sound very grumpy, but we are negative on the asset class outright, but also on a relative basis relative to global equities.
Markets Are Too Pessimistic on Domestic Cyclicals
Karen Gilchrist: Finally, is there anywhere you feel the market is overly pessimistic where you’re feeling more positive?
Sebastian Raedler: Yeah, that’s very important, and you’ve given me a great chance not to sound like a complete sourpuss. Cyclicals versus defenses in Europe are at a 30-year high, reflecting this extraordinary level of optimism. However, some cyclical pockets have completely failed to participate in this historic rally.
These are specifically the domestic cyclicals—that’s small versus large caps. It’s a very cyclical trade at a ten-year low, even as cyclicals overall are at a 30-year high. This is things like German equities, which have been big underperformers, partly because they had the completely wrong sector mix for this environment. They are overweight chemicals, overweight autos, overweight software. That’s a toxic mix. We think that cyclically things are improving in Europe, as we have discussed.
Secondly, we still think that the German fiscal delivery is on its way and will come, even if people have lost patience with that. We think at the index level, all of this is priced given these stretch metrics. But for the domestic beaten-up cyclicals that have not participated, there is a window of opportunity to catch up with the positive sentiment priced into the rest of the market. And I think it illustrates the point that I made earlier. If you have bad news priced in, the current environment is a great opportunity for that to be reversed. There’s simply no bad news priced at the index level, so I think you should be cautious. But these things that have been forgotten, where no good news is priced in—even as a slightly more cautious and bearish person, I would say there are opportunities there, because they are priced for something significantly more negative than the rest of the market.
Karen Gilchrist: Sebastian Raedler, thank you for your time. For Morningstar, I’m Karen Gilchrist.
