Morgan Stanley's Michael Wilson on AI's Next Market Winners

0:00Michael Wilson: The midterm elections are now behind us and Wall Street is weighing what the results mean for the economy and investment markets. But beyond those political headlines, AI continues to reshape the economic outlook and raise questions about where the next wave of value will be created. So joining us today to make sense of all of that is one of Wall Street's most closely watched market strategists, Michael Wilson, Chief Investment Officer at Morgan Stanley. It's Saturday, November 7th, and this is the AI Proving Ground podcast from Worldwide Technology. Let's jump in.

0:38Michael Wilson: Michael, welcome to the Ad Brewing Ground podcast. How are you? I'm doing great. Thanks for having me. Yeah, I'm excited to have you here. We are 48 hours after the midterm elections. And I know I've seen you say that the quote unquote blue wave was not surprising, not to you, certainly not to the markets. It was kind of already baked in. But I have seen you say what is interesting and what is potentially a risk is what happens in the following days, the following weeks and months. From a technology perspective, what did you mean by that? Yeah. Well, first of all, I would say that the magnitude of the blue wave was not expected.

1:09Okay. So let's talk about a couple of specific races, which is, I think it's really weighing on like the semiconductor complex today and some of the data center, you know, direct plays. It's really what happened in Texas and Ohio, right? So these are two states that were vying for data center projects. And, you know, the folks who won those elections are sort of anti-data center, at least for now. Okay. We'll see how long that, you know, persists because, you know, I have a general view that, you know, money talks and there's a lot of money behind this and there's a lot of initiative and, you know, politics are, you know, are politics.

1:42And so people are going to run on a platform. I'm not saying they're going to go completely back on their promises, but, you know, they may be, they may reconsider, hey, this is good for our communities. You know, Let's analyze this. Maybe let's up our asking price. What are the other benefits we can receive here? So I do think that's a legitimate concern, given those two elections in particular, that there could be a slowdown now for a couple months, three months. But I don't think this is going to stop the freight train of the capital spending. And the main reason is because they've already raised the money.

2:13And in some ways, it could actually balance it out because we have a shortage on both power and chips and labor and other constraints. And so by slowing things down a bit, you can actually prevent some excesses, meaning inflation. This stuff is very inflationary. So maybe if you slow down the path of that spending, it can balance itself out a little bit better. And stock market is already pricing this pretty quickly.

2:38Michael Wilson: Yeah. I mean, even beyond data centers, what about, you know, there's so much complication going on right now with what's happening, ups, downs. I mean, certainly we're talking about big providers like the NVIDIAs of the world, or we're talking about NeoCloud providers. I mean, how do you expect or how do you anticipate kind of the rest of the market to react beyond just the data center boom? Yeah, well, I mean, politics, you know, they say elections have consequences. Okay. And I just want to remind the listeners that, you know, almost every midterm election, there's a change of power. Because generally, we're never happy with our, you know, leaders in Washington.

3:11So there's, it's an opportunity to kind of vote and send a message back to the policymakers, effectively. So like I said, this one was a little bit stronger than I think the blue wave was initially expected to be. For example, I thought it'd be 52 seats, maybe 51 or 52 seats. It's more like 54 plus. So that's a bit more of a mandate if you want to go down that road. You know, the president still has veto power. It's going to be hard to really change legislation in a persistent manner. I do think it may expedite a solution in Iran because that message is clear. That's what people are most unhappy with is the inflationary consequences of the war.

3:49And also Americans are sick and tired of these wars. And that was known. This is where I think the president, you know, probably presses luck potentially. He's a rational guy when it comes to politics. He understands this is a loss. So I'm hopeful that this expedites perhaps a peaceful solution. And so one of the things I've been thinking about for next year, and this has nothing to do with the election, but it could be related, is, you know, the Strait of Hormuz in general is this choke point that's existed for a long time. And now with it being exposed, everyone in the world is asking themselves, wait, why did this exist?

4:22Why did we have this choke point? So if you're a leader in a country that's energy insecure, and there are a lot of them, you can't allow this to happen again. So I think what could end up happening is there's going to be a massive supply response, more drilling activity, perhaps in regions that were apprehensive to do drilling or building different routes out of the Middle East through pipelines, which is now ongoing. And we could probably, a year from now or six months from now, be in a situation where we have a glut of both not only crude oil, we already have a glut of crude oil, but also product if there's refineries built.

4:55So I'm trying to always think six months in advance because that's what stocks do. Stocks trade in the future. So if you're reading me the headlines today, chances are we've discounted 80 % of it, 90 % of it. The last couple of days is a little bit different. But generally speaking, the market's trying to figure out what's going to be happening in six months. And I think lower oil prices, lower energy prices could be one of those.

5:19Michael Wilson: So speaking more broadly now, beyond the midterm, six months from now, what is the market telling us about how it's valuing AI? Because that is certainly all over the place, too. Yeah. So all capital spending cycles go through multiple phases. And we could talk about Internet 1.0. We could talk about the fracking, the CapEx cycle. And what typically happens is people get to – it's a real thing. These are all real things. Like, this is pretty exciting. So they get excited. Then they raise a bunch of capital to build it out because it's like, yeah, we want to fund that. This is going to definitely have a return on capital.

5:55And so the early winners typically are the beneficiaries of that capital spend. So in this case, semiconductors or power turbines for the energy production. And then what happens is you get to a point where the rate of change, the second derivative of that growth, peaks out. That's what happened in June, meaning the peak rate of change on semiconductor growth peaked. So the semiconductor stocks started to underperform. Some of the power players started to underperform. And what started working was software stocks and some of the adopters of this technology, like non-tech companies even, the people who are going to benefit from this spend.

6:33So we're moving into that second phase. is we call it going from the enabler companies, okay, the ones building it out, to the adopters. And if you think about the hyperscalers, it's interesting to me, the hyperscalers are both. They're like the tweener, right? They're the enabler. They're building these LLMs, but they're also the biggest adopters. If you look at the most efficient users of AI so far, it's the big hyperscalers because they're technology companies. They know how to use this technology to then make their core businesses better. So that's why if you look at the hyperscalers who kind of own the stack, they don't just provide the compute and the LLM, but they know how to use it to sell into their core business compared to, say, a neocloud business, which is a commodity compute provider.

7:19Well, you can see why there's been a divergence now between those two. And then what's really exciting about these capital spending cycles is when it diffuses into the broader economy and you see completely new businesses created, they didn't even exist. And they transform our lives. Like that's when the average person, you and I will feel in our daily lives.

7:41Michael Wilson: Yeah. I am interested in that adopter versus builder mentality. So you're saying, you know, there's a general shift that will be taking or perhaps is already taking place in which the market will value companies that are more efficient with AI or putting AI to better use as opposed to just the builders? Or are you saying maybe potentially in place of builders? Well, it's both. So I would say on the builder side, as is typical, the initial euphoria is everybody's a winner. Like if you're in that bucket, everybody wins. But that's not really right because there are monopoly profit type businesses that benefit from that.

8:15And then there are ones who are kind of on the outside who kind of got bit up because they just kind of ran with the pack. So what's really good, what's positive to me about this correction since June is the market has separated what I call the higher quality enablers versus the lower quality ones who are no longer being funded as much because their commodity provider or they're really not, their part of the stack is not as important. Okay, so that's one part of it. And then the market now is looking across the investment environment saying, okay, well, let's try to figure out who's going to use this stuff.

8:50And then, like I said earlier, software companies, The infrastructure layer. I think there's a really, I mean, I've been doing technology for 35 years as an investor. And in every cycle, okay, the real value add is always in the application layer. It's always in the software, right? Because that's the interface that makes technology easy to use for guys like me and you. You might be a little more tech savvy. Maybe you code. I don't code, okay? But if I see an application that's like user-friendly, now all of a sudden I become a consumer of that application. and then that uses compute. So that's where we are now.

9:24We're building that interface, the killer app.

9:27Michael Wilson: I was just going to ask, so are we on the search right now for a killer app or is it potentially, I mean, we're still very early in this whole process and I know a lot of these companies take years and years and years to build. So are we close to having a killer app? I mean, that's a crystal ball question, but like, should we expect it soon? Have we seen it already? We've seen attempts at it already. You could say ChatGPT or Claude or Muse now. I mean, these are killer apps. OK, but it's that is that are they big enough and are they specific enough to really diffuse like the adoption rate is still quite low, even within the enterprise?

10:00You know, like I see it in our firm, you know, there's a little bit of apprehension still. It's not easy for people. It's like that. I don't really like prompting is a skill, you know, and it's different. Like, say, the killer app to me for Internet was email. I'm older than you. So I remember when email came out, I was like, what is this is incredible. And it's pretty intuitive, right? I'm writing a letter, basically. You know, some are longer than I'm sure. But I mean, like, you're basically communicating to this thing. And wait, I don't have to fax things anymore? So, like, that was just so simple.

10:31Or a browser, a web browser. And then, of course, then you have applications on top of that. Once you get access to the internet, then you need applications. So we're early days on that. I could imagine vertical solutions for tax, for travel, for legal services. These things haven't been built yet, but they're going to be built. I mean, they're going to be built probably pretty quickly. And so is somebody going to own that whole stack? Probably not. I think it's probably going to be these – like certain companies are going to dominate certain verticals. And so it's coming.

11:06Michael Wilson: Yeah. So where do you think we are in the cycle right now? Now I've heard you compare this to kind of the late, late 90s when a little bit of the, you know, the credit markets were wobbly kind of in the early stages of the dot-com era. So why is that comp, you know, in your mind? Well, I think, first of all, we have seen a lot of build out already. We'll do a trillion dollars plus of capital spending, which, you know, even GDP adjusted. That's on par with what we were kind of doing in the late 90s for the telecom build out, the enterprise build out, and the consumer build out for the infrastructure.

11:37So that's one aspect. The second aspect is when these things get going is before everybody gets funded. You know, money is basically free. It's like, well, you're AI. Here you go. How much do you need? Let me write the check. And now what the market is saying is like, well, you know, we'll write you a check, but it's going to cost you more now. So that's the market doing a very good job of differentiating and saying your cost of capital is going to be X and yours is going to be Y based on what we think, you know, we're like the potential payoff. and the risk of that particular bet. That really started about 97, 98 in that particular boom, meaning you had weaker credits start to underperform and spreads really widened.

12:20And that was the beginning of a four-year credit bear market. 1998 ran all the way until 2002. But as you probably know from looking at past markets, the equity market didn't peak until early, mid-2000. Now, is it going to be perfectly fit to that? Who knows? Okay. But two years ago or a year and a half ago, there was a lot of angst. You may remember this. Like literally, it was just getting going. People were already worrying about a bubble. And I remember telling some folks internally next to them, I was like, guys, we haven't even started raising debt capital yet. This is all being funded by cash.

12:56You know, so like until they start borrowing debt, borrowing money to pay for this, you know, I'm not concerned. Now we have borrowed a lot of debt. So you should be watching. You need to be, if you're an equity or debt investor, you need to be a bit more selective. And that's exactly what the stock market and the credit markets have done since June, is they're being a bit more prescriptive on credit versus equity, or I'm sorry, quality within credit and equity.

13:23Michael Wilson: That's good. I mean, so the capital is getting smarter here. Should the enterprise leader kind of follow that same game plan? You have to be smarter about where you're placing your bets? I think so. I mean, and also understand your customer needs. You know, I think in these initial phase, what we're seeing now is a lot of enterprises jumped on board because nobody wants to fall behind. So everybody says, well, I'm doing AI too because you're getting rewarded to say that. And let's be honest. I mean, most organizations are not set up to move that fast. The other thing we're finding is it's very expensive, you know, and security and safety, right?

13:56you let these agents kind of in-house, and we've seen a couple of examples where they kind of run amok, or they go rogue and they cost you tens or hundreds of thousands of dollars that you weren't budgeting for. And then, of course, the security challenge. So I think a lot of companies are now thinking about this defensively. They're spending money on AI to protect their customers, for example. Like if you have a big install base of customers, you don't want to get hacked. So you're spending a lot of money on AI to protect you from AI intrusion. So yeah, it's going through. I've seen two examples where the spending is really aggressive.

14:36One is the tech-savvy large enterprise who has resources. And by the way, we're in a positive earning cycle. So there is money to spend right now. And then we also have what I would call the startup community, VC. If you're a small company, you can basically do a lot of functions now on AI, marketing functions, legal functions, accounting functions, and essentially much cheaper. And they're spending a lot of money over there. So the middle part is not spending a lot of money yet on AI. So we're not penetrated. I mean, we got a long way to go on penetration.

15:12Michael Wilson: Yeah. I want to get back a little bit to the bubble talk. We're talking about the telco buildout. We're talking about the dot-com boom. I mean, certainly that ended up busting, although there were certainly plenty of winners. And look at where we're at today in transformational technology. But what makes this era different, or is it not different? And we're just going to end up, yeah, there's going to be organizations and companies that fall by the wayside, and there's going to be clear winners. But at the end of the day, 20 years from now, we're going to be in a transformational state. I think there's like two or three major – first of all, it's not different.

15:45Okay, every capital spending cycle ends with malinvestment, and the early spenders of that will lose. But the beneficiaries of that are the people who didn't pay for it. So it's a really good example from the internet boom. We had all the telco companies, the C-Lex, the cable companies. They spent all this money on connectivity and building the World Wide Web, and it's very expensive. Well, I mean, the biggest winners of that were the companies we now know and love, right? It's the big four hyperscalers. Well, they were tiny, ex-Microsoft, but they were tiny. They wouldn't even know who they were because they got to build businesses on top of this basically internet highway for free.

16:30Marginal cost was zero. And I don't see how that's going to be any different. Now, the difference this time is the big spenders this time are also pretty technologically savvy. They have good core businesses. They have the full stack. So that's why the hyperscalers, I don't necessarily think they're going to go the way of the telco companies. Some might. But because they have good core businesses, they can actually, like I said before, they're enablers. They're spending a lot of money on this. They're going to probably not make the return they think they are on the base technology. But they can use that technology to then make their core businesses even a bigger moat and create new businesses, whether it be in the healthcare vertical or the educational vertical or the manufacturing vertical.

17:13And we haven't even gotten to physical AI yet, which is like a whole other avenue.

17:17Michael Wilson: Right. I assume by, you know, when you say way of the telco, you're meaning kind of commoditization there. how can organizations, enterprise leaders guard against retaining some of that or as much of the value generated by AI so that it's not all going towards the hyperscalers or those model builders? Well, once again, you got to be an adopter and you got to fit and you got to be creative. I mean, one thing about AI, I've been fairly active in using it myself. I mean, everything speeds up. I mean, you can get so much more done. It gives me then the ability to be more creative and think about, oh, well, how can we build a better mousetrap?

17:57And I'm in a pretty narrow world, right? I'm in an investment community, okay? But think if you're an artist. Think if you're a scientist and you can discover the next compound. Think if you want to build the next educational facility. We all know that secondary education in this country has gotten ridiculously expensive. I mean, it's stupid. People can't afford to go to school now, to colleges. Because they have too much, even if you're successful and you graduate a good degree, I mean, you're carrying$300 ,000 worth of debt. It's nuts. And why is that? Because they built these huge facilities, basically monuments, okay, and essentially country clubs for like dorm rooms and stuff.

18:35It's kind of nuts and stadiums. And it's like, what are we doing? And we're trying to educate people, aren't we? So you can imagine like a completely retooled educational system, which by the way, is much more democratic. You know, you don't have to go through the application process. says, hey, I want to get a Harvard-level education for one 100th of the cost. I'm not in it for the, yeah, I want some social experience too, but I don't want to pay 100 grand a year. So just think about the efficiency of that. You can raise a level of education across the entire population. It's incredible.

19:08Michael Wilson: Back to the markets right now, you're talking about how investors are looking into the future. They're looking at a six-month horizon, if not a little bit longer or shorter. What are investors rewarding organizations for right now? And how is that starting to evolve? It's a great question. So we spend an inordinate amount of time, even before AI was, but now we can do it really quickly, which is, but we can do, we can basically ask that question. Okay. And there's things called factors in the market. So for example, you know, a quality factor is like, so your balance sheet is you're not over levered.

19:45Okay. or you don't have like crazy operational leverage, okay? That would be like a good sign for quality. One that we focused on a lot a year ago was CapEx to sales, which is a low quality factor, meaning you spend more CapEx to your sales rate. And it was being rewarded from basically April of 2025 until about June of this year. And we were saying, okay, these companies are being rewarded for spending more money, which is unusual. That only happens coming out of a recession, which I think is what happened in 2025. and that all flipped on a dime in June, meaning all of a sudden, if you said, hey, we're going to spend more money on CapEx, the stock went down.

20:23And so now what the market is saying is, we're good with you spending money, but not recklessly. So the market is rewarding operational efficiency. It's rewarding quality factors. It's actually punishing you for doing the opposite of that. Growth is still favored over value factor. Size, the market is now valuing size and scale versus small caps, which was not the case between April 25 and June of 26. So yeah, it's already evolved. And that was the call to be making in June, which we did. And now it's sort of like now quality is sort of well known and it's crowded. So we're looking for the next one.

21:06And I think we're about to transition, not back to low quality, but we're going to transition into what I would call benefit. It's going to be different factors that benefit from the adoption. So we're looking for things that would signal companies who are actually taking advantage of this technology. Yeah.

21:22Michael Wilson: Moving from that builder to adopter type of stage. You've mentioned June a number of times now. Is that 25 or 2026? This year, 2026. June, 2026. What happened in June, 2026 that has your attention? Three things happened that we were fortunate to recognize kind of real time, because you want to do it before the market figures out, was, first of all, we had to set the stage. Coming off of April of 2025, that was a very low point for what we call earnings revision breadth, meaning the rate of change on revisions for companies was negative. So most companies were kind of revising numbers down. And that's a signal that You're at a trough from an economic momentum standpoint.

22:04It was the bottom of a recession, effectively. So coming out of that, you typically see then that second derivative, the rate of change, accelerate, which it did. But it ultimately reaches a point where it's unsustainably high. So in June, what we noticed was, particularly for semiconductors, it was at an all-time high in terms of that rate of change. It's unsustainable. And once again, markets trade in the second derivative. So that was the big one. It was that the rate of change on earnings revisions and earnings revision breadth for what I call early cycle groups like semiconductors, financials, some of the consumer areas was peaking out.

22:40The second thing that was going on is we had a new Fed chair. Kevin Warsh came in in the spring, and he took over in May, June. And we noticed there that he's not as fond of using the balance sheet to print money. He's talked about that. And in fact, he did slow down the balance sheet expansion starting in June. So the rate of change on money supply and liquidity for the market started to roll over. Not negative growth, but less growth. And that was the second thing. So those two were the main two things that changed. So Fed kind of moving from being overly dovish to a little more hawkish money supply growth, and then a rate of change and revision breadth.

23:21Those are classic signals that you're going from what I call an early cycle recovery to a mid-cycle recovery, which leads to quality factor outperforming.

23:31Michael Wilson: Okay, so that's leading to quality. Investors are now looking to reward more of that adoptive phase. You mentioned reckless spending. It is interesting right now, and I'm certainly not an economist, but what's going on right now, you're seeing just massive, mind-boggling numbers in terms of what some of these companies are spending on infrastructure. We always hear that the demand is there, but they still are getting dinged a little bit in the market. Is that what you mean by where we're at? Well, it's sorting it out. I mean, I would even say within the hyperscalers, they haven't traded together for the last year and a half, meaning two are doing really well and two of the stocks are not trading well, and then they've morphed back and forth.

24:12So if, for example, Google was the big winner last year in the stock market, and this year now it's meta, because they kind of leapfrog each other back and forth. So it's not as black and white as either in or out of the whole thing. Once again, the market's being more discerning. It wants to own the one where you're getting accelerative growth. And they can say, oh, this is now the horse I want to ride. And I'm sure that's going to change again. Amazon is being punished because it's a consumer business as much as it is a hyperscaler in AWS. And the consumer has been under pressure because of oil prices, higher interest rates, which I think is going to change now.

24:50I think you're going to see rates down, oil down, which means maybe Amazon is the one that takes the lead into 2027 because it's got that consumer business.

25:00Michael Wilson: We're running short on time, and this has been excellent insight. I want to kind of just ground it back to our audience, which is enterprise IT leaders, enterprise AI leaders looking to just make the best decision that they can the next day on Monday. And it's such a rapidly moving landscape, it's hard to keep pace. So given everything that we've talked about and everything you're monitoring, what are some of the decisions that enterprise leaders, specifically within the realm of AI, need to start making soon so that they can position themselves for whatever's coming next? I think you have to embrace it.

25:33I think there's still people who are kind of in denial that this is coming and somehow I can avoid this in my industry. It's coming for every industry. And that includes the physical industry, too, ultimately with robotics. So you need to get on board. And by the way, most well-run companies are doing that. But what that really is is a cultural revolution within your business. You need to reward your employees for being adopters. And so that's point number one. Point number two, though, I think it's just as important, is you need to be cost efficient. One thing we know about technology is that marginal cost goes down over time.

26:09So it's a balancing act. We want to be fast adopters, but we also don't want to break the bank because if you're paying for the highest end model, trying to do all this stuff, letting agents run rogue or, you know, it can get very expensive. So it's a balancing act in that regard. And then, you know, look over your shoulder. You know, talk to make sure you're in the weeds with the private world because that's where I think a lot of the stuff is happening in these smaller organizations where they got nothing to lose. And they can sneak up on you pretty quick.

26:42Michael Wilson: And I know I said that was my last question. I'm going to add one more on top of it because you reminded me, you said, you know, we're not even talking yet about the physical world. And you said that was going to be a major inflection point. So what, you know, what do you anticipate happening once, you know, we start seeing robotics out there or world models or digital twins, whatever it might be? Well, I think the biggest question is when does this technology go from inflationary to deflationary? Okay. Which is typical, by the way, of all technologies. By the way, I'm going to take a sidebar because this is a tech audience.

27:10Let's talk about fracking. Okay. because it's the same thing. I mean, and most people don't think about this, but if you think about what fracking was, it was spending a bunch of capital looking for oil in reservoirs that were harder to tap. And they used technology to get at that oil. And as a residual, they got natural gas, but there are no pipelines to cap it. So they have just flared off. And so we have all this excess natural gas now in these regions, be it South Dakota, Pennsylvania, even Ohio, and of course, West Texas. And that's an example where the marginal cost of natural gas is now negative.

27:46They can't give it away. So, you know, like there's real, it's very deflationary. But it was very inflationary when they were, it's kind of similar to data centers. Like these communities just kind of popped up out of nowhere where you had, and it was interesting because you remember you had the housing boom. So you had all these workers who were skilled workers in the housing who were out of a job because the housing thing blew up. But then they could easily transfer some of those skills into the oil industry, physical labor, drilling for oil. And so I see a similar setup today where it's inflationary for labor.

28:18We're building all this stuff, but ultimately it's going to be deflationary. So I think that's when it becomes physical and you're really getting free, not free, but much cheaper labor, it becomes deflationary. And that has two consequences. One, it's really good for the user. okay it's a dangerous thing for economy that's over levered the worst thing for an over levered economy is deflation right because you have to pay back your debt in the same dollars i think we're a ways away from that i think that's like a story for like three four five years from now but it's to me that's going to be the most important development from when we can actually do this a physical ai it's going to be a deflationary event particularly for labor

29:03Michael Wilson: So you're saying potentially that we may be underestimating exactly how much AI we'll be using over the next decade. And to go along with the adopter versus builder, the winners will be, what are you doing with the oil? What are you doing with the AI? That's right. And I agree with the statement. I hear it all the time. There's an insatiable demand for intelligence. Yeah. But there's not an insatiable demand for intelligence at a certain price. So every technology or capital investment cycle, the marginal price goes to zero once again. And eventually, the marginal price will go through the marginal cost because you probably overpaid for some of that capacity in the rush to build it.

29:44And by the way, if the marginal price doesn't go down to the marginal cost, it's not going to work. It's not going to be affordable to the average user. But then when it does get there, that's when the really exciting stuff happens. that's when you get free intelligence. Imagine free intelligence. And there's enough people in the world who are willing to take that free intelligence and do some pretty cool stuff. Yeah.

30:06Michael Wilson: Yeah. Sky's the limit. Well, Michael, you've been gracious with your time. Thank you so much for joining us. And we look forward to seeing what happens next. You bet. Me too. All right. Thanks.