Full Transcript: Adam Parker on AI Revenue, Software, and Markets
Trivariate Research Founder on Semis, SaaS, Healthcare and What Value Investing Gets Wrong
Justin: Adam, thanks for joining us and welcome back to Excess Returns.
Adam: Thanks for having me.
Justin: So much for sell in May and go away, I guess.
Adam: I always think those things are a little rough. We actually did some analysis on that point, and you need like a few hundred more years to prove it’s statistically significant. So it’s one of those things where you just could have had some spurious things that happen and then people think it’s a trade. But yeah, I don’t-- I’m not a big believer. And I think we studied it’s the third worst month out of the twelve in the last hundred years, but it’s nowhere near statistically significant. So I’m not a huge believer in those. That’s a little astrological.
Justin: Yeah. Well, I kinda said it tongue in cheek because I know that at Trivariant, where you’re focused on is analysis over anecdotes, combining quant and judgment, and I think understanding what is different between cyclical and structural in the markets. And one of the things that you said recently, I don’t know if it was from that Barron’s article or it might have been some other piece of research that you guys put out, but one of the things that you said is the stock market leads the economy, not the other way around. And that you’d rather be wrong based on your inputs rather than listening to some economist forecast. So I just thought that might be practically a good place to start in sort of articulating how you think about the markets and build your view.
Adam: Yeah, I’m not-- I didn’t mean to insult anyone with that. I mean, the idea, like when you work at a big firm, I used to work at Morgan Stanley, and the big firms have these year ahead and quarterly outlooks, right? And you just have to imagine, whether it’s Goldman or JPMorgan or UBS or whatever, you have a lot of really smart people that they have disparate jobs around the world, right? Interest rate strategy, economics, fixed income, equity strategy, currency, whatever. And so the idea of having some firm-wide outlook, the way it worked at Morgan Stanley, and I assume it’s similar everywhere else, is there’s some meeting and it’s a couple of hours. And when I was at Morgan Stanley, there were forty-four economists, and they would tell you their outlook for the economy. And then like a few days later, the interest rate and currency and credit guys would tell you, “Okay, given the economic assumptions, here’s what we assume will happen to the currencies and the rates or whatever.” And then the last meeting, maybe two weeks after the original one, was the equity strategists were supposed to tell you what you thought of the US equity market. But ostensibly you’re supposed to use as inputs the economic view and the currency and everything else. So like every one of those meetings, almost like a broken record, I always thought we could definitely all be wrong, but there’s no chance we’re all right. You know what I mean? Like that’s... And so I just, yeah, I totally just thought I should just come up with my view because some of these things are really hard to forecast, like currency and interest rates for sure, and oil and other things. So we might as well just use our judgment. And I think what you said is definitely demonstrably true, which is the stock market leads the economic data, not the other way around anyway. So the economist meeting should be after the equity
Justin: meeting. Do it in reverse, I guess.
Adam: Yeah, totally. And also like economic skill, which I think when we were younger, and I’m older than you guys, but like when I was younger, it was considered like a good degree, like scientific and, you know. But now I don’t really think there’s a lot of skill to that degree because, you know, Gemini or Claude or whatever, like you can get all the economic data immediately, the forecasts. Like one of my criticisms of economists is I don’t think economists even know what already happened because the data gets revised or the definitions aren’t exactly what people think, and so we feel like it makes more sense to just use the information in stock price action to predict what’s gonna happen.
Justin: What do you think of the idea? I think one of the things that investors are trying to figure out now is whether or not we are in sort of a stock market bubble. I think a lot of investors are looking back to like the late ‘90s and 2000s and trying to draw correlaries between today’s market and back then. What do you think of the question around whether or not the market is in a bubble or not?
Adam: I don’t love that phrase. I think it’s really hard to time when the top is. What I would tell people is like, look, a lot of the smartest people with the best access to information and the most high-powered computers are spending all day long trying to make this call. And so the idea that you’re gonna sort of sell exactly at the top of your memory exposure or your semi exposure, I mean, I’m not God. Like that’s a tough call, right? So I would rather try to figure out how to neutralize that exposure, find other AI revenue, diversified stuff from... You know? So, but I don’t think it’s a bubble in terms of price action. There are some things that remind me or rhyme. So I’ll give you a couple thoughts. One is, I feel like every time there’s been a top, two things have been in place: hubris and debt. And so if you look at, like, say, Sam Altman at OpenAI, lots of arrogance, tons of debt, little bit worrisome, right? I could see somebody saying these kind of giga cap or whatever they’re called, IPOs that are coming are a little kinda toppy, right? The total dollars in capital spending from the hyperscalers, like the capital spending dollar growth, that’s definitely kinda toppy. But in terms of the price action, like we could still have a huge move from here to get it anywhere near where we were on valuations and even return moves versus the TMT crisis that peaked in March 2000. So I don’t know. I don’t even know what I would do with that information anyway, because I don’t think I can time the peak, and I feel like I’d rather be a couple weeks late and ride it over the edge than miss out on the continued upside.
Justin: And do you think that that is sort of what the market is sniffing out here? I mean, it’s been very resilient in the face of the war, higher oil prices. Consumer sentiment is in the tank here. What is the market-- Is it sort of seeing the earnings and the fundamentals through all that? Is that why the market’s sort of marching higher?
Adam: Two thoughts on that. I mean, it’s easy to say, “Oh, the market’s not trading on fundamentals. It’s retail idiots or crypto bros,” or whatever people wanna say on the institutional side when they’re underperforming. But the truth is, it might be trading on fundamentals. It might be trading on a distribution of outcomes of 2030 fundamentals or 2031 or something in the future. And by the way, just actually interestingly, and this is maybe just coincidental, but the stock market S&P is up around nine percent or so YTD as we’re recording this. The bottom-up estimates for this year are about eight percent higher today than they were at the beginning of the year. If you look at the sector level, the sectors where the estimates are up the most YTD are tech and energy. They’re also the two best performing sectors. The sectors where the estimates are down the most are consumer discretionary, financials, and healthcare. They’re also the three worst performing sectors. So there are some things where the fundamentals have tied to performance, but I get the concept of the stocks are ahead of the fundamentals for the AI kinda data center build-out. I mean, there’s no doubt about that. And so I think people are just trying to say, maybe unknowingly, “Hey, I’m gonna call that top before everyone else. I’m gonna get it right.” And I think that’s a tough, tough, tough call. I mean, every day you’ll read some doomsday thing saying this is the top or whatever, but the market’s up 100% since the first time I read one of those.
Justin: Do you have any thoughts on when the market is going to start demanding or wanting to see returns on this CapEx build-out?
Adam: Yeah. I mean, look, we did some really detailed work recently, Justin. I mean, I think that’s like a whole year of question. We think only nine percent, two hundred and sixty-two out of the top three thousand public US equities are currently generating any meaningful AI revenue. Around sixteen percent of companies are saying on the cost side that they are doing something or will be benefiting from it. So part of the answer to your question is, and this is like a bit of the innocent until proven guilty mode the equity market’s been in, is this is all still in front of us, man. So like getting bearish early could be the challenge. You know, maybe it’s gonna come at ‘27, ‘28, et cetera. The earnings estimates this year bottom-up are low twenties. People think it’s mid-teens again next year. So you’re starting to see some of it on the revenue and cost side. But let’s unpack your question a little bit more. I mean, what could derail the rally would be some implementation issues, right? Like what if a big company says we have to run things in parallel or maybe there’s a whole host of new... I had a hedge fund guy tell me last week they have to hire an agentic safety officer, right? So you could start seeing like less productivity in terms of like revenue per employee or something that people thought because other jobs surfaced. I mean, I think that’s possible, and so there could be some sort of implementation delay. But I guess the short answer is in the next year, man, we better see some serious productivity from the companies.
Justin: To your point about that, you know, hiring that new person, Bezos was interviewed on CNBC, and he was actually making the argument that AI is gonna create more jobs, which is total non-consensus. I agree.
Adam: I didn’t know that, but man, I’ll take it whenever somebody like that and I can accidentally stumble on the same thing somebody like that’s saying. I agree. In fact, I wrote about that in my most recent piece, that I think there’s a weird consensus view that’s the opposite of history. So history would show every new technology creates a whole host of new jobs. Some get displaced and others get built. But somehow everyone thinks the opposite. You know, everyone’s worried their kids are gonna be unable to get employed because AI’s gonna take all the jobs. But I’m not sure. I actually wrote this in my Sunday note. I didn’t take brain science or whatever to do it. I just literally typed into Gemini, “Can you suggest for me jobs that might be formed because of AI?” And a whole bunch of stuff came up in finance and law and healthcare and startups and marketing and corporate. All kinds of jobs. And I started to think, well, maybe this will be kind of more of a hiring boom, both corporate-wise and economically, than people think.
Jack: Do you think-- I’ve been working on this thesis myself. Like I think maybe this is gonna be maybe the most transformative technology we’ve ever seen. So to your point, there’s gonna be tons of jobs we can’t even think of. But also because it’s so transformative, like we might have a little more short-term pain than we have with other technologies as well. So like maybe a little bit more in the short term, but more long-term benefit. Is that fair?
Adam: It’s always possible. I think the current data, if you look at like new hires for undergrads and masters, really do support your view. There’s definitely a slowdown a little bit now. So that’s possible. But I’m just starting to hear about, think through, and see more AI-related jobs forming. And people with no skills in these areas can get up to speed pretty quickly, right? You can watch a one-hour Claude coding thing on X for free and get your skills up. So I’m optimistic that there’ll be a lot of opportunities even in areas that people think will be obsolete, like the law. I think there’s gonna be more legal issues around AI, not less.
Jack: You mentioned the IPOs that are coming in. I just wanna ask you before I forget about it, like do you have any thoughts at a market-wide level if that matters? Like a lot of people are talking right now about the fact that we’re gonna have multiple trillion dollar IPOs coming in-- where does that money come from and all that. Do you think that matters to the market overall that these IPOs are coming?
Adam: So it’s tricky. I get asked this question a lot. I think it’s a good question. I’ll just give a couple of thoughts and then let you react, and then let’s sandpaper our block, okay?
Jack: All right.
Adam: So for most of my career, I’ve been hamstrung by compliance issues on trading larger cap US equities. So periodically, I’ve trafficked in the illiquid small cap, microcaps. So there’s been a time where I bought ten thousand dollars worth of a fifty million market cap company, and that trade moved the market cap to seventy million. So my ten thousand dollars turned into twenty million. That’s thought one. Thought two is, when I worked at Morgan Stanley, their call on interest rates were wrong every single year I worked there. And one of the main reasons is because they would always say there’s a lot of supply coming online. All these bonds are gonna be issued. It won’t be met by demand, and so therefore bond yields will back up. And we probably all know interest rate people who have called zero of the last fifty bond-breaking, you know, breaking the financial system calls, right? So what happens is, when people are afraid, they buy the ten-year anyway, and so you don’t get pro rata the same amount of volume in up days as down days, right? I’m just giving you a couple random thoughts. So I’m not sure. I think what possible mistake people are making in their head is like, “Oh, there’s gonna be a trade and a half at the IPO of SpaceX,” and so that trade and a half is just gonna come straight out of the other Mag Seven, and they’re all gonna go down.
Like, that’s not the correct math, I can assure you. But I do think there will be some pro rata selling because a lot of passive or tight tracking error long-onlys have to -- can’t be underweight something that’s big. And as you know, at least with SpaceX, it looks like it’s gonna be in the S&P pretty much right away, and so it’s gonna be an index problem for long only guys. So let’s say they do seventy-five billion at one and a half trillion. Is that what it is? So it needs to have a big long only. I mean, Musk is choosing basically who gets the shares, right? So no big long only is gonna get the pro rata amount. Let’s say that’s one and a half, which is what, three percent weight in the S&P. So I think there’s gonna be a lot of forced buying for a really long time, and I wouldn’t be shocked if this thing went to two and a half trillion or three trillion just almost independent of price, just because people have to buy some to catch up and not much is floated. I think it’ll be different with OpenAI and Anthropic just because they’re not gonna be in the index day one, and so there’ll be more time.
But I don’t think that necessarily means the money pro rata comes out of like the other Mag Seven. I think it probably smooths across multiple other names and/or some new capital gets allocated to it. And then the other thought I guess is, a lot of institutional investors, you know, one thing we do for a living is people send us their portfolios to do custom risk work. And I think a lot of institutional investors I analyze their portfolios for are actively underweight Tesla. And there’s a general view that, oh, maybe it’ll come from Tesla in particular in some sort of Musk allocation. And I worry about that. Like I feel like that could be dangerous. I feel like I wanna be close to market weight Tesla, not underweight, because I could see some combination-- what if SpaceX buys Tesla? Like something that could happen that could really screw you if you’re underweight. So anyway, those are my thoughts, but let me pause again.
Jack: Well, first of all, a lot of people do think those companies are gonna combine. Like a lot of people in the tech community think that that’s sort of the inevitability at some point.
Adam: Yeah. So but yeah, I don’t know how to answer that, but I think if you’re strictly long only and really have low tracking error, you’re gonna have to... And you’re 100% allocated, you’ll probably just pro rata sell a little bit of everything to buy some up, you know?
Jack: Yeah, to your point, I mean, I think it’s impossible to know what’s gonna happen, but I think the mechanics are just so interesting because this is just the largest IPO ever. And so you think both on the passive side, I mean, the passive funds are kinda falling all over each other right now to make sure they get this thing in as quickly as possible. And then you’re also gonna have, like, on the active side, you’ve got some funds that hold it now that are gonna be in violation of their mandates because they’re gonna hold too much once it becomes public. So I’m just thinking about the dynamics and all the re-- It’s just interesting. I think I don’t think there’s really much of a takeaway from it. It’s just interesting to think about everything that’s gonna have to happen mechanically once this comes out.
Adam: I agree, and I’m not sure. I don’t know. All I know is that people who get paid a lot of money to do this stuff for a living in other examples that I pointed to are consistently wrong at
Jack: it. That’s right.
Adam: That’s
Jack: true in a lot of things in investing, right?
Adam: Yeah. But I know I don’t know, so I’m honest about it.
Jack: Which is the best way to be.
Adam: Yeah. I mean, look, I think the biggest issue is they’re just floating very little, and the people who need to get big chunks, three, four percent positions, like if you’re these big ten trillion club kind of asset managers, like you’re gonna own-- 3% is your bench weight or whatever it is. Like, you have to own, and you’re not gonna get any shares at the IPO. So they’re gonna be buying this thing almost like with no instruction until they can get a lot of it.
Jack: I wanted to talk about software, but first I wanted to talk a little bit about factors because I’m kind of a quant nerd myself, and I think it was a Barron’s piece I was reading. You talked about gross margin expansion is probably one of the most powerful factors you use. Can you explain why that is?
Adam: Well, there’s a couple reasons. One, I think the more down the income statement you go, the more BS there is in there. So I remember years ago, United Technologies, which was run by a guy named George David, he was the CEO. And this is before they had the Raytheon transaction and spun off Carrier and Otis. But when he retired, he said, “I’m proud that we beat earnings fifty-nine consecutive quarters.” And I thought to myself, WTF? Like, how do you beat fifty-nine? Like, are the analysts uniquely stupid in this sector? No. Right? So, are there cyclical businesses you’re in? Of course, they’re in helicopters, air conditioners, elevators. So it only just proves that there’s all these levers you can pull the farther down the income statement to beat the numbers, right? Like sure, I’m sure economically, there were some quarters where they beat it and they operated well. I’m not saying there wasn’t a lot of that. But you can’t beat fifty-nine consecutive quarters, which is basically fifteen years, without having a lot of wiggle room. So answer one is you wanna be toward the top of the income statement, not the bottom, okay? And answer two is more than any other margin level, the change in gross margin and the change in multiples, EV to forecasted sales or price to earnings, are highly correlated. So businesses that have higher gross margin trade at higher EV to forecasted sales. It’s not necessarily true at the earnings level because of perception about over or under-earning or other stuff. So I think it’s that combination. I think as an analyst also, you can spend meaningful time getting a disparate view versus consensus on gross margin, right? Because if you think about gross margin, there’s the revenue. The best thing to get high gross margins is raise pricing without any loss in unit demand, right? If I just charge more-- If I charge more per widget and still sell the same number of widgets, that’s really high incremental margin, right? So we do a ton of work on this, and there’s all kinds of other things too, input costs on the cost side as well, depreciation, labor, material. So I think it’s because it’s more predictable or analyzable and the change is correlated to the multiple.
Jack: I wanted to ask that because software has been basically the poster child for high gross margins for a long time, and now we have a lot of question-
Adam: But it’s changed up. It’s changed up level.
Jack: That makes sense.
Adam: Yeah.
Jack: They’ve consistently had-- And if they print eighty, they don’t improve consistently high gross margins. Right. But I know you’re not positive on software right now. And it’s such an interesting thing to think about because they’ve been such really good businesses, like a lot of quality funds have them. But you’ve got this thing out in the future that’s leading everybody to question the terminal value of these businesses. And I was wondering if you could... You had said in, I think it was the Barron’s piece, “My highest conviction call right now is not being sucked into any rally in software.” So I’m wondering if you could talk that through.
Adam: Yeah. I mean, so we’ve had this view for a couple years. We called it our North Star to like semis over software. Software and semis currently have the lowest correlation they’ve had since AI started. So at the very moment we’re recording this, I think there’s some notion that software’s actually gonna be defensive, almost like a consumer staple relative to semis. And there could be some truth to that because the near-term P&L volatility is gonna be lower, right? You sign up ServiceNow, you’re not getting rid of the contract this year, et cetera, et cetera. But to take the tenor of your question, why have I been negative on software? Well, when I see this much multiple contraction like we’ve seen, the probability of the earnings miss is high and ultimately, eventually, the sales miss. So here’s why I think the median software company will underperform by a lot from here. When you look at the analyst estimates, they’re for eighty percent gross margin and pretty high net margin every year going forward with modest, very modest revenue deceleration, twenty, nineteen, eighteen, that kind of stuff. So somebody could say, “Well, these are really high growing businesses. They’re high margin. What the hell? It’s oversold, SaaSpocalypse,” all that crap. But the software companies usually work when the revenue is accelerating. So in order to get revenue acceleration, how am I gonna have the same gross margin and net margin every year going forward? I’m probably gonna have to invest money on AI tools, OpEx and CapEx, et cetera. And so I feel like phase one is they’re gonna have to invest to make sure they retain the business. Phase two is the big customers, the chief technology officers from JP Morgan and Morgan Stanley and Goldman, they’re gonna push back on pricing because they know that their position relative to the software companies is strengthened. So that’s why I think they miss on earnings. It may be in the next five or six quarters, not immediately. And then eventually they’re gonna miss on sales because companies are gonna be able to, in some cases, attach their own internal tools or other stuff. What the analysts do is they call the CTOs. They all have a few contacts. One of the biggest BS phrases on Wall Street is channel checking, by the way. But what they’ll do is, “Oh, I do my channel checking.” They have like one friend who’s a CTO, and that guy tells them, “Oh, we’re never getting rid of ServiceNow or Intuit or whatever.” And then they call the IR guy from the company, and they cry in their beer together about how it’s oversold. But that doesn’t tell you anything about the 2030 distribution of outcomes. So that’s kinda my general thesis. I’ll just caveat it by saying, look, there’s gonna be some software companies that work that you can’t get rid of, and I think they’re gonna be the ones, and I have high conviction on this, that are more expensive and growing faster. The last thing I wanna do is buy a cheap software company that doesn’t grow and try to play it for its cash flow. Like, the market’s right on average to make the ones that are cheaper, cheaper because they’re more likely to be made obsolete. The ones that are more expensive, software like in security and other areas, which we pointed out CrowdStrike, Palo Alto, others, which have been pretty good stocks recently. The reason is because you’re just gonna continue to pay them. Like, Justin, if you worked at a big bank, and you’re like, “Hey, I quad-coded security, and it’s just as good as Palo Alto. Let’s save the money.” Let’s say there’s a breach two years from now. Not only are you fired, the CEO’s fired for being a dipshit for listening to you in the first place. So they’re just gonna pay the software companies in certain areas as if they’re consulting firms just to kinda CYA almost. So I think there’ll be some software winners, but they’re not gonna have the same pricing power, and they’re not gonna have the same margin profile in five years as they have today. So that
Jack: it’s interesting because everybody-- when you see something like getting killed like this, there’s always the tendency, everybody wants to buy it. That’s just kind of a natural reaction. But I thought what was good about what you’re talking about is, like, a lot of times when you get these big declines in things, there is like some sort of major structural impairment. There’s
Adam: information.
Jack: You know?
Adam: Yeah.
Jack: Yeah. Yeah, it means something, right?
Adam: Yeah. I mean, if you go back to the Nasdaq, you know, March 2000 to October 2002, 77.4% decline, there were still 10 rallies of 15% or more during that two and a half year decline. So you can get some pretty powerful countertrend rallies, and people will say, “Oh, it’s oversold,” and... But I’m just trying to get the 77% direction right. I’m not gonna be smart enough to get these little kind of powerful countertrend moves. So if you wanna buy software because you think it’s defensive against an AI semi trade, buy the expensive ones that are growing faster. Okay. And so you’re gonna own a few software companies in an S&P index fund, but my North Star is still semis over software for sure.
Jack: So you’re still a believer. I mean, we’ve had quite a run up in semis here. You still think they’re pretty attractive?
Adam: I mean, my current view is sorta like market weight semis, underweight software. Okay. But I like that relative trade. I mean, I think in semis it’s trickier because, when you take that AI revenue story I told you a minute ago where 9% of the companies seem to be having some AI revenue exposure, there are about six different categories, right? Like, there’s memory and semi-cap, there’s vertical and edge, there’s infrastructure, there’s power. Like, there’s different buckets there, so you gotta start transitioning where you think they’re over or under-earning more, where more of the acceleration is in front of you than behind you, you know?
Jack: Yeah, I don’t know if you agree. I was listening to-- You listen to that Gavin Baker interview with Patrick O’Shaughnessy recently?
Adam: I’ve known Gavin for a really long time from when I used to be a semiconductor analyst, so I’ve listened to him. He’s a very compelling communicator.
Jack: Yeah, he’s very good. But he was just making the argument that the semi-cap equipment stuff is very, very expensive relative to like the DRAM stuff. And like, kind of both situations can’t be true right now. Like, there might be some relative pricing issues within that. I don’t know if you agree with that.
Adam: Yeah, I mean, no, I like that logic. In fact, what inspired me to do this AI revenue note I talked about was actually Micron and Caterpillar. So Micron at the time I wrote the note was trading at seven times earnings, next year’s earnings, I think maybe six times buy-side expectations, and Cat was trading at 30 times. And I sort of feel like they both can’t be true. Like, if you’re saying one’s massively over-earning because of a data center build-out inevitability, and the other one is not over-earning, those seem incongruous. I will say that there are people now I know that are trying hard to find cheaper DRAMs and sort of say, “Well, I don’t need to buy expensive ones if I’m not using them.” So there’s gonna slowly be a convergence between demand and supply growth. So it makes sense to me Cat’s more expensive than Micron in absolute terms. I think the debate is just, is the chasm too wide?
And it’s hard to sell stuff when it’s got this much momentum and prices going up and whatever. I mean, the memory stuff’s really interesting. If you look back, even down chain with Sandisk and other stuff, there may be people who are forced to buy the shares just because they’re too underweight, right? Like the risk management point you made about when we were talking about SpaceX a minute ago. So I kinda feel like when momentum’s this good and fundamentals this good and a lot of people are underweight, you may still see the shares act better than people think.
Jack: I wonder, taking gross margins and like carrying them to the market overall, I’m wondering, you talked about how you think gross margins might be coming down across like a median company.
Adam: Yeah.
Jack: And I think that made you a little bit cautious. Can you talk about that?
Adam: Yeah, I mean, I was just making this point, like two things can be true at the same time. Like, we could have the most high margin market cap ever. I think over forty percent of the S&P 500 market cap has more than sixty percent margin. But we could also have issues where the median company’s margins aren’t going up. And part of that was just because of input costs with commodities and oil and other stuff, and less pricing power, which we’ve seen across multiple consumer areas where businesses are bumping up into pricing issues where they just can’t pass it on at the rate they were. So the median company, take the top five hundred, so the two hundred and fiftieth biggest company, their gross margins are actually about a hundred and fifty bps lower than they were 18 months ago. So you’ve seen that trend where it’s been harder for the average company to take pricing. But if you take the total dollars in the market, gross profit dollars, and divide it by the total, and that’s because of Nvidia and Micron and other businesses that are currently generating a ton of high margin business. I mean, forty-five percent of the entire S&P 500 earnings growth year over year in Q1 was from Micron and Nvidia.
Jack: And this actually gets to a question I was gonna ask you later but I’ll ask you now, which is this idea that, earlier this year we had all the stuff that was not working, it started working. Like, the median company was outperforming, value, international stocks-- it was all going crazy, and then we had this war and like a complete reversal in all of that. We’re back to where we started. Like, I’m wondering, do you have any thoughts on that? Like, is this kind of we’re back to this is the way we’re gonna be now?
Adam: Well, they’re just growing faster. I mean, I think one of the thoughts was, it’s funny, I said it to you earlier, like, well, it’s change, not level. But it’s funny, like, two months ago, and actually this was just kind of well-timed and lucky, right in the 20th of March I wrote a note saying, “Well, maybe level does matter, just, like, the level is high enough.” If tech earnings go 45% or whatever they would expect to do at that time for this year, another 25 next year, it’s not like tech’s gonna be 60% cheaper in a year. It’s just gonna go up, right? And that’s kind of what’s happened. Like, the level was so high, the growth was so powerful that it could handle multiple contractions. So the market could still be okay even if the median company has gross margin pressure because the actual growth’s okay. So we think growth grows low teens this year, maybe 10% next year, so the market can trend higher even if there’s a little bit of multiple contraction, and that’s kind of been my view. But generally level change is better than level. Like, something improving is better than buying a stock where margins go from one to five is better than buying one where they go from 80 to 76 or whatever.
Jack: Yeah. One of the things you’ve been ahead of, and we talked about this in our last interview with you, is you’ve been avoiding like these common value metrics everybody’s using in terms of thinking all these things are cheap based on these standard valuation metrics.
Adam: Yeah. Valuation doesn’t work to pick stocks. I mean, I don’t know how many times you have to-- how many years you need to know that. I actually have kind of changed my view on that to the point of, I just almost think it’s arrogant now if you say, “I buy a stock ‘cause it’s cheap.” You’re just saying you think the optical valuation has information in it. You see it’s cheap, but no one else with all the computer power in the world and all the people that are looking at it think it’s cheap. What the hell are you talking about? Like, stocks that are cheap are cheap for a reason on average. The reason you wanna buy a stock that’s low is if the estimates are too low, not if the price-to-earnings is low. Like-
Jack: Right. It’s about expectations versus reality, right?
Adam: Yeah. Low-
Jack: Like, a lot of those stocks are cheap because expectations are close to reality, which is bad.
Adam: Buy low, sell high means buy where the estimates are too low and sell where the estimates are too high, not the price-to-earnings.
Jack: On that estimates point, that’s another thing you mentioned is this idea that the penalty for missing has become very, very severe recently relative to the reward for beating. And I think you also mentioned that in the cheap companies, that’s even worse. Like those companies are getting killed, right? Is that right?
Adam: Yeah. So maybe not everyone knows this, but about 70% of companies beat estimates. So you gotta unanchor your head from it being 50/50. And that’s either because of that thing I mentioned earlier where there’s wiggle room in the P&L, or they guide conservatively, or the analysts aren’t what they believe or whatever it is, right? But like 70% beat. So what’s really interesting is stocks that just got more expensive over the last quarter, they have a higher probability of beating estimates. Stocks that just got cheaper have a higher probability of missing. So there’s information in the change in multiple, right? And then the probability you beat a second time, given you beat the first time, is higher than the unconditional probability. So all of a sudden it’s not just momentum why the stock’s up. There’s information in it, like the market’s on average correctly predicting that the company’s gonna beat estimates. And so we’re in this regime right now where, like you mentioned, the penalty for missing is way harsher than the reward for beating, but the market’s up because more companies beat than miss. And so the wise anchor advice I give to some of my clients is just don’t own stuff that misses. You know what I mean? But it used to be you could say, “Hey, I know current conditions are kinda weak for this business, but I feel like the stock’s cheap enough and it’s down a lot, so it’s kinda discounting a recession and I believe it’ll be in better shape in two years and I’ll buy it here.” And like that was a completely reasonable thing to say in the 1990s and even first half of this millennium. But like now the market is crushing the companies that are cheap that miss because it’s saying, “Hey, there’s a serial correlation there. The probability they miss again is higher than the unconditional probability.” So I think that’s changed. Like the market’s gotten increasingly anticipatory and right on average.
Jack: Is that something that’s regime-specific or is that something that’s just continued to change over time? Like more and more companies get penalized over time as we learn more about this and the market gets smarter about it.
Adam: Yeah. I always struggle a little bit with that question, like giving a confident answer because how long do I need to see this before I conclude it’s not spurious but it’s not just regime or cyclical? We wrote a note last week actually on like 10 things I used to believe that I don’t believe anymore, and they were like things that really have changed since COVID. So more like five or six years, which I think is long enough to say it’s not spurious. And I think, I can’t prove this, but I think the way money’s being run now, where so much money went to passive or super low fee, and so much money went to multi-strats that are running factor neutral with high turnover, that I could see this being more structural and mechanical in how the market behaves. So things like valuation not working or high quality not working or other things probably are more structural.
Jack: That’s one of the things that you said early in the interview too, and I think it’s such an important lesson for investors, this idea of like the N is very important in these variables. Like, you’ll see on X all the time, someone will be like, “Every time this happens, the market goes up X percent or something.” And you’re like, “What’s your sample size?” Like five times or something like that. It’s so important to remember that. That’s such an important point.
Adam: I don’t know this guy, but I want to. There’s a website called Tyler Viggen. I get no financial benefit from mentioning this. T-Y-L-E-R V-I-G-G-E-N. And I bought his book also, and it’s called... It’s a little too far for me to reach without putting my midsection into the camera here, which nobody wants. But it’s tylerviggen.com, and it’s a spurious correlation website. And it’s just got all these awesome things of like, you know, it just shows, you know, like a small sample size, like nine or ten data points, annual data points, and it’ll be like a curve, and it’ll be like number of rainy days in Maine and the number of times, you know. It’s just completely ridiculous spurious correlations, and they look perfect. And it’s sort of always these good examples to realize, like you could be concluding something that’s preposterous with a small sample size. So I like that. I agree with what you said there. The only challenge is like, does this get spread around enough on X where it becomes like a self-fulfilling prophecy? And like that we don’t know, and it could be, right? Like the May comment you made, Justin, at the beginning, like maybe people just start thinking that September’s always a crap month, and then you get a couple first bad days at the beginning, and it’s like self-fulfilling, you know. And then you see it posted on X by a thousand bots and dudes and girls like-
Jack: It’s funny because in the quant world that made me think of ESG because ESG is something you should not get a premium for over time, like doing something good. But at the point that everybody started adopting ESG, like ESG went crazy. Like again, the idea is once everybody starts believing ESG is gonna get them extra returns, like it kind of went nuts. And it might be a bad example, but I just thought about that when you were talking about that.
Adam: Yeah, I mean, that one’s a tricky one because it’s confounded so much by size and quality, right? Like bigger businesses can afford to have a lot more ESG frameworks or whatever. I’ve actually spent a lot of time on that topic. I wrote a note, I think in 2021 called-- we analyzed the ESG ETFs that were out there at the time, and I think the note was called forty-nine percent S&P, forty-nine percent NASDAQ, two percent ESG, ‘cause that’s basically what the stocks in the ETFs were. So it’s trickier. I mean, I think we can all agree that G matters, like just in governance, for sure. But it’s harder to measure some of the S and E stuff, you know? Like is-- are you guys E? I don’t know. Like... But whatever, I mean. You know, like it’s hard to measure, like what do you mean?
Jack: Yeah, yeah. No, definitely.
Adam: Yeah.
Jack: And obviously that does kinda die down anyway. Like that was all over everything in the news for a long time, and now you don’t ever hear about it anymore.
Adam: Yeah. In Europe you do. I’m on the board of a European public company, and it’s definitely bigger in Europe and Scandinavia. And I think also some allocators want that, and so obviously if you’re trying to gather assets, you may have to show that you are. But I think in the end, like your point up front is right, you wanna-- it should be meritocratic, like whoever generates excess returns, to use your phrase, like should command the premium.
Jack: Just one more thing on estimates. You had this thing I saw in one of the interviews you were talking about, this idea of estimate achievability. And I know that goes beyond what a lot of people are asking, like “Is it a beat or is it a miss on this quarter?” And that goes beyond that. So can you just explain that?
Adam: Yeah. I mean, I guess tying into the thought of like the penalty for missing is harsher than the reward for beating, like how do I tighten things up? One thing I like to look at, and this is gonna be on the nerdier side of what we’ve talked about so far, which-
Jack: We like
Adam: nerdy. I already worried I sent this on the email.
Jack: Nerdy is great for our audience.
Adam: Okay. But I do a regression for every stock between change in sales and change in income. That regression, if you do like a year over year scatterplot, is the incremental margin. So let me just make sure we’re all on the same page. Let’s say I’m a thirty percent gross margin business, right? So my gross profit is thirty cents on a dollar. The question is, what’s my gross profit gonna be on every new dollar of revenue above today’s level, right? So I do thirty cents on my first dollar, but maybe I do forty cents on my second dollar. So if I get two dollars in revenue right now, I’m gonna have seventy cents in gross profit or thirty-five percent gross margin, right? So we spend a lot of time thinking about what is embedded in the consensus estimates for incremental gross margin. And if the incremental gross margin expectations are in line with the business model average, that might increase the probability of achievability. If they’re way above what the business usually does, then something has to fundamentally tie. Did they just move all their employees to a low-cost jurisdiction? Did they have a big drawdown on depreciation roll-off? Did they have a big price increase that everyone’s excited about their new product? There has to be something that you can explain to me why the expectations for incremental margin are way above what they normally do. Because analysts are decent at-- Actually, analysts are not very good at anything, but they’re pretty good at knowing fixed and variable costs from the businesses. And so you should have a decent idea about estimate achievability by looking at incremental margin. So I think about that a decent amount, actually.
Jack: Yeah, that’s an interesting point because the idea is as it scales, those margins are gonna change. So a lot of people are kinda looking at where it is now, but looking at it as it scales, how it changes is probably far more important to think
Adam: about. Yeah, when I used to do semis, the classic way to look at it is what are the incremental gross margins? And I am guilty of applying semiconductor logic to a lot of sectors, but look, we think about it in our own business too, right? When you’re scaling, you have costs that come in the door, but your incremental margin should be pretty high.
Justin: I just saw that Micron just crossed one trillion dollar market cap. It’s up sixteen percent today.
Adam: Yeah. You never would’ve guessed that based on twenty-five years ago, the business they were in. So yeah, they’re the thirteenth biggest company now or whatever the heck they are. It’s crazy.
Justin: What are your thoughts on sort of the Mag Seven right now?
Adam: I had a financial advisor tell me last week that they got ten new clients that are like ten million or above Micron shareholders that were just like regular employees and now have ten million bucks in Micron stock.
Justin: Oh, geez.
Adam: So it’s been a boon for the financial advisor industry.
Justin: There’s gonna be a lot of new SpaceX millionaires out there too pretty soon, so yeah, financial advisors are gonna be busy.
Adam: Yeah.
Justin: I wanted just to get your thoughts on the Mag Seven here. I mean, you have most of these companies, Microsoft, Meta, Google, Amazon, they’re all spending ridiculous amounts on this AI infrastructure build-out. But then you have Apple, basically the lone wolf that is basically doing it through partnerships and not necessarily a big CapEx spend. So I just wanted to get your thoughts on how you’re thinking of positioning the Mag Seven and anything company-specific that you’re paying attention to.
Adam: Yeah. So our general advice has been to tell institutional investors to keep market weight the bolus of the eight names, and we count Broadcom in there, too. The idea being that, one, they’re pretty macro stocks, meaning they don’t have a lot of company-specific risk. Market goes up, growth beats value, large beats small. You can kinda tell how they did that day. Two, they’re the most well-covered securities in the world. So I always think it’s funny where somebody thinks they have a non-consensus view on one of these names. That’s pretty hard to have. And then three, it’s a really concentrated market, so I don’t think people should have their highest conviction views on one of these names. They’re better off being close to market weight of the group and then having higher conviction in other things that are a little bit easier to have an edge on. So most institutional investors have big active weights there or they’re underweight, and retail’s been a little bit more overweight, and that’s how the weight’s the weight. So I feel like the answer is keep close to market weight of the group. But if you put a gun to my head, I kinda like NVIDIA still, just as I feel like the hyperscaler spending is not gonna wane anytime soon, and their position with CUDA and all the money they lost for years, it’s gonna be hard to do any complex modeling without that platform. So I don’t love how the stock acts when they blow out their quarters, but I just don’t see how they’re not gonna grow so fast for the next couple years.
Justin: Adam, we always love having you on with these conversations. And we could talk for a lot longer and about a lot of different things. But I wanted just to sort of close out on some of the work and the findings that you’ve done, which I find very interesting, in regards to spinoffs. And I think your research shows that the average spinoff tends to meaningfully outperform its industry or peer group for a couple of years, while oftentimes the parent of that spinoff lags. So can you just sort of explain that data that you’re seeing in spinoffs?
Adam: Yeah. I mean, I agree with that. I think spincos have unlocked a lot of value. It’s complicated, Justin, because oftentimes the remainco is getting a tax benefit, and that’s the catalyst for doing it. But I could see a regime here in the next few years where AI productivity, smaller companies being able to compete with bigger ones, creates a wave of spinoffs actually. And I can tell you it’s somewhat popular at the law firm clients we have and on board discussions to consider spinco. So I think it’s a pretty good way to unlock value, particularly when the spin is in a slightly different sub-industry than the remainco, where maybe there’s not as many synergies, et cetera. So it’ll be interesting to see if that’s the new wave, because M&A generally hasn’t been that effective. The acquirer underperforms. So I think, you know, there’s always... You could make an argument that we got so many big companies now that you need to sort of make them smaller again, and maybe that’ll be cyclically the next wave that happens over the next five or ten years.
Justin: Before we let you go, what is the highest conviction out-of-consensus idea that you and the firm has right now?
Adam: I mean, out of consensus and wrong are really correlated. And so for me, it’s been the healthcare sector that I think will work. I just think the whole point of AI, or a big point of it, is we live longer and we’re more productive while we’re alive. And so we have this aging demographic where people demand services, tools, diagnostics, drugs, et cetera. These businesses tend to be low margin, probably can benefit from incremental efficiency, can benefit from predicting their customer employee behavior. And so I feel like the market’s telling me the government isn’t gonna pay for stuff at a rate that I think is wrong. I mean, there’s no political will on either side to cut the spending the amount that’s in some of these names. And so I tend to like the drug distributors, the managed care, the Quest and LabCorp, like the tools, like all those kind of businesses that I just think will end up growing above GDP and have margin expansion. But that has been out of consensus/wrong as an aggregate call. Like the drug distributors are great stocks, but then they get sold off harshly and managed care works. And like we haven’t seen things work at the same time, and I think they should. So I like healthcare a lot. It’s uncorrelated to the tech bet which we’re making. It’s the most number of transactions in any sector. You know, top three on pipes and IPOs and follow-ons and all that stuff. And so I think portfolio construction-wise it works too. But it’s been out of consensus/wrong.
Jack: I also think, though, by the way, if there’s something you wanna be really optimistic about with AI, like healthcare is a place. Like we may see some incredible breakthroughs in the next decade or so.
Adam: Look, I spent a lot of time this year. I took an AI for healthcare course at MIT a couple years ago, and I got sort of mesmerized by the potential. Do you guys have wearables, any, or Oura or any of those things?
Justin: Jack?
Jack: I wear
Justin: a Whoop. I wear a Whoop. Yeah.
Jack: Yeah.
Adam: Whoop or whatever. Okay. Yep. So I think what’s gonna happen in a few years is every house that’s built is gonna have a giant box that’s like a wearable by your bed that’s gonna be tracking the electromagnetic waves and everything. And it’s gonna send you an email saying, “Hey Justin, we made you an appointment. We noticed you’re drinking too much Gatorade during your, uh...” Okay. But it’s gonna know like everything that’s wrong with you and really like the efficiency, like the healthcare tech convergence is gonna be massive, and I think it’s underappreciated.
Justin: Thank you very much, Adam. That’s been wrong.
Adam: That’s been wrong. That’s the most out of consensus call we have. And I feel like what the market’s telling me is 0% chance healthcare is the best performing sector in the market over the next five years. And I think it’s thirty or forty percent or whatever. I’m just trying to play that arb.
Justin: Thank you, Adam. We really appreciate you coming back on and joining us. It’s the day after Memorial Day, hitting the morning with a nice conversation so we appreciate it.
Adam: Thanks. Thanks for having me. Great to see you guys.