In this episode of Monthly Macro, macro analyst Richard de Chazal and William Blair’s co-head of technology, media, and communications research, Jason Ader, unpack the sharp selloff in software stocks and the growing fear around AI-driven disruption. They explore why the “software apocalypse” narrative may be overblown, how AI could ultimately expand the software market, and what investors should watch for as winners and losers begin to separate.
Subscribe here: Spotify | Apple Podcasts | Amazon Audible
Podcast Transcript
00:05, Chris
Hi everybody. It's Wednesday, February 26th, 2026. Welcome back to another episode of Monthly Macro. This month markets have delivered a few surprises. None bigger than what we've seen in the software space where many long-established mega-cap names have sold off sharply in a relatively short period of time. So, joining me to discuss, we've got macro analyst Richard de Chazal, as well as William Blair co-head of the technology, communications and media research sector Jason Ader.
They will unpack what's really driving this move, what's changed, what investors are going to be missing as the AI narrative continue to evolve and so forth. So, Jason, Richard, thank you both for joining. Richard, I'll let you take it from here when you kick it off.
01:08, Richard
Great. Thanks, Chris. Jason, thanks so much for coming on to the call.
I'm sure you're very busy, at this time, known as. Not only is it earnings season, but, certainly with what people are calling “SAP Apocalypse” or the software slump. I'm sure it hasn't been particularly fun times.
So, maybe just, you know, as Chris said, unpack it a bit. I mean, what do you think is going on? What has suddenly changed, you know, is there one single catalyst or is it sort of a bunch of things, you know? Did investors, kind of, all herd to one side of the boat and now they're sort of flipping to the other? What do you think the deal is here?
01:45, Jason
Okay, sure. Thanks for having me, Richard, Chris. I think a lot of this, sort of, inflected around the Christmas holiday period where, you know, honestly, people just had time to experiment with this new technology, especially the new anthropic model, Claude Opus 4.6. And it just was, you know, a big leap forward, I think, in terms of functionality, not just around, you know, the ability to automate coding, but actually complete tasks.
So, this idea of a gigantic, I think, has been talked about for a couple of years. But, I think there was a sense around Christmas time and then moving into January, as you know, things like Claude Cowork came out and other plug ins that we were entering, kind of, the next stage in AI. So, the initial stage was chatbots with ChatGPT, and now we're moving into this, agentic deck period.
And I think that just freaked a lot of software investors out. You know, a lot of the pundits were saying, well, you could just vibe code software, you know, just tell Claude what you want to build and it's going to build it for you. And you don't need any expertise in software engineering, and you don't need to buy off the shelf software.
I mean, it's a bit silly, and obviously a lot more complicated than that, but I think that was, sort of, the generic fear, very much amorphous, you know, boogeyman kind of situation. But I think it just legitimately raised concerns on the durability of software moats, you know. And don't forget, software has always been viewed as, like, a super predictable sector, a lot of recurring revenue, mostly recurring revenue, high margins, low churn.
And this explains, in some ways, why private equity has been so enamored with the sector. Valuations were high in terms of, like, enterprise value to sales, enterprise value to free cash flow, historically. So, I think this was, sort of, a reset on valuation in a lot of ways because people start to think about, well, you know, are these cash flows, you know, actually, as repeatable, as durable as we expected, or as we thought.
And, so now we're just going through this discovery process. And, you know, I think we will see more discrimination, you know, between winners and losers versus, you know, everybody in the bathtub. So, I think that's right. Now, where we are, I think it's being driven very much by just fear of the unknown.
04:23, Richard
What do you think of the idea that I'm sort of looking at this and thinking, okay, well, maybe, I mean, some people are just painting, kind of, this horrific story that software is dead, and, you know, we've reached the end of the road there. And now we're going from clicks back to bricks, if you will, or, you know, I think that's probably wrong.
But what I do think is maybe investors have realized, to a certain extent, to get the AI future that they, kind of, foresee, there's that, sort of, you know, wants to scale exponentially versus the reality that we still need all the bricks and mortar to, kind of, get us towards that future. And maybe investors are rotating from some of these software stocks into kind of the more industrial materials, energy side of things, which you kind of need as the next vehicle to take us to that future. Do you think that's part of it, at least?
05:25, Jason
Yeah, absolutely. I think there's a, like, pretty, secular sector rotation going on where investors just feel a lot more confidence in the predictability of the revenues and the cash flows of those infrastructure plays versus, you know, the software companies where, again, there's just a, you know, a lot of change. Ultimately, if you want to get a return off all this bricks and mortar, all these data centers, you're going to need software to, you know, be monetized.
Applications, you know, are going to be what drives the monetization of all this capital investment. So, you know, at a high level, we think software is, you know, far from dead. If anything, we think AI really grows the software pie. The question is, you know, who are the winners and losers? you know, within that pie.
And clearly some of these private AI labs are going to be major winners on the software side. And so, you know, who gets impacted by that? Who gets disrupted? And, I think that's what people are trying to figure out right now. And really nobody knows. So, you know, when you have that kind of a situation, I think people just, sort of, shoot first and ask questions later.
But, ultimately, like I said, you do need to have some monetization vehicle for all this investment. Otherwise, you're never going to get a return on it. I think we've, you know, started to see some concerns, you know, clearly over the last year on just the magnitude of the capital investment. And, so, some of the big, you know, data center builders, you know, have also taken a hit, as people start to question, you know, the return profile on these investments.
07:12, Richard
What do you think, I mean, it strikes me that a lot of this is narrative driven at the moment. Like, I mean, we're just going through earnings season. I think they've had pretty good earnings. There hasn't been, you know, much weakness on the fundamentals versus the narrative here. And, you know, this is, sort of, the reports, the scare reports that are, kind of, being put out or like oh, this is, you know, I'm in 2030 now and writing this report, about what the future looks like there.
But, do you think the market is maybe underestimating how long this adoption takes place? And, you know, to what extent it actually takes place? There's, sort of, a mismatch going on there.
07:55, Jason
Yeah, you know, like the quote that sticks with me, that, you know, a lot of the tech, you know, executives have thrown out there, which is, you know, a lot of times people overestimate the impact in the near term and underestimate the impact in the long term on a platform shift like this. This is clearly like the biggest platform shift, you know, probably we've seen in the tech space, I mean, I think it's bigger than the internet.
Internet, you know, ultimately was about distribution. You know, this is, like, a cognitive, a massive cognitive change, which, you know, really does raise questions on the future of knowledge work and knowledge workers. So, I understand, in a lot of ways, you know, why there is so much, kind of, consternation right now.
And, yeah, you've seen a lot of these, sort of, viral blogs and reports that are, you know, sort of, doomsday scenarios. But I agree that, you know, you have physical constraints. I mean, particularly, on the energy power side, on the, you know, the physical infrastructure like chips and memory.
I think politics is going to play a big role here, too, in the midterms and in 2028 election. And just in terms of like, you know, NIMBY data center pushback. And then just the larger, sort of, societal questions on, sort of, is this a good thing for our society? If, you know, 50% of knowledge workers lose their jobs in the next five years, as some of the, you know, AI folks have forecasted. So, I think this is going to take a while to play out, but obviously investors and the stock market are, kind of, forward looking animals.
And I think, yes, as I said, I mean, this is a massive paradigm shift. And, you really can't disprove the, kind of, bearish narratives yet. And so, one of the frameworks that we've adopted is just, okay, which software companies are likely to accelerate revenue growth in 2026? If your revenue growth is accelerating, you know, maybe in the future you could be impacted by AI, but you're certainly not being impacted in 2026 if your revenue growth is accelerating.
So, I think investors are going to become a little more discriminating as we move forward. And, as you noted, this earnings period has generally been pretty strong. You know, if you look across the software sector and then, you know, the outlook, I think, is going to be the, you know, where the rubber meets the road over the next couple of years in how the numbers play out.
And are companies seeing AI as a tailwind, right? If you can, you know, if you can say, all right, our revenue has accelerated because of AI, you know, you're probably going to get a better multiple, and you're going to be viewed, you know, as more of a winner. And so that’s going to be hard to deny for the, you know, kind of, the bears out there.
10:58, Richard
So, is that how you're, sort of, dividing up the winners and losers? Like, you know, what kind of criteria?
11:04, Jason
I think that's the best tool that we have right now. I mean, I think this is fluid and obviously there are certain other elements here that we think about, you know. I think when you think about, like, the type of model that you have, pricing model you have, like, for instance, I think consumption-based models are probably safer, you know, kind of more tied to outcomes and value versus subscription models and seat-based, especially seat-based subscription models.
But I think seat-based models will need to evolve into something that has a usage based or, kind of, outcome-based component. But, again, that's, sort of, back to the fears of, you know, the number of knowledge workers. If we see shrinkage in knowledge workers, that obviously is not good for seat-based software models, which is the end market that they serve, as knowledge workers.
It's, sort of, almost industry by industry, too. A lot of those, sort of, vertical software is where we're seeing a lot of the new startups and a lot of the disruption. So, there's a lot of really exciting private companies that have emerged. But, I think the incumbents, the existing companies, you know, as we've written about, if the marginal cost of delivering new software has basically gone to zero because of this, you know, this AI revolution, that should also benefit incumbent players that can build software and features a lot more quickly across a lot of different, sort of, adjacencies.
So, and they have the customer already and they have the data already. And so, there are, I think there's a very strong, sort of, counterpoint argument to, you know, the fears out there on vibe coding and, sort of, you know, lower barriers to entry on just like the code itself.
12:57, Richard
Isn't the reality that, sort of, software costs as a share of total tech costs are very low. And, you know, tech costs as a share of total, sort of, operating expenses are quite low, as well. So, I mean, how sensitive are companies to that? I mean, if software or SaaS is, I think it's, I read, you know, 5% of total tech costs, which tech costs are, sort of, 10% to 15% of total operating. I mean, you know-
13:30, Jason
Is it worth it? Yeah. Yeah. I've heard similar numbers. I think I've heard software is, something like 8% to 12% of total tech spending. And tech is obviously that, sort of, rising. And over the last 20 years, every year tech rises as a percentage of total corporate spend. I think that will continue.
Look, ultimately software, like, let's face it, software eats into labor. I mean, that's what software is about. It's about, you know, efficiency, automation. This, sort of, is like, you know, putting that on steroids. So, yeah, I do think there will be an impact on, sort of, the number of people that you need. But I also very strongly believe in Jevons Paradox, which basically says if you lower the cost of something, you're going to see, you know, much more consumption of that something.
So that's what I mean by, like, the pie for software, I think, is going to grow a lot larger. And that, I think, underpins, sort of, my confidence in the sector. But look, whenever you have a platform shift, there are winners and losers. There are literally thousands. I mean, I don't know if I'd say tens of thousands at this point, but certainly thousands of startups, AI native startups that are well-funded, have a lot of smart people, you know, have a clean sheet of paper, so there's no, sort of, technical debt or, you know, kind of, an existing product that you need to protect, you know, your, sort of, classic innovator's dilemma. So, certainly, I expect a lot of these startups to be successful. A lot of them are going to fail to, as we saw in the dotcom era. And then the, you know, the AI labs, you know, the big private AI labs, I mean, are crushing it. I mean, they're innovating at such a rapid pace. I think that's what, again, is creating some of this is anxiety because it's just a moving target.
15:15, Richard
Do you have any idea where, sort of, adoption rates are of AI? I try and look at some of the data, you know, it's all over the shop, really. But, my impression is still relatively low, sort of, 10%, 15% ish?
15:28, Jason
Well, on the consumer side, that's probably right. It's still very early but growing rapidly. It will, you know, gradually build as, you know, more people, have exposure to it. You know, OpenAI is going to be introducing ads. You remember in, like, the early days of the internet they said, you know, we'll never have ads on the internet?
So, I mean, I think, you know, obviously you got, you know, OpenAI has something like 900 million users and, you know, 95% don't pay a cent for it. So, they're going to be monetizing that more so. And, you know, the Super Bowl, you know, making fun of OpenAI for ads was a great ad. But it was, kind of, you know, a bit ridiculous because, you know that that's the future. I mean, there are going to be ads with these chatbots.
But, on the enterprise side, I think adoption is really low. I think that's where, you know, some folks, kind of, overestimate, sort of, the timeline, you know, for AI adoption. It's very hard for enterprises to adopt such a, kind of, groundbreaking, you know, organizational shifting technology.
It's not just because you’ve got to train people on that, you know, change the workflow and the, sort of, process, but it's also just fears on data security, data governance. You know, sort of, risk management that you think about an enterprise has to really abide by. And this stuff is really new.
And, you know, you probably heard some of the horror stories on some of these new tools and how it exposes, you know, personal data or whatever. So, I think it's just going to take time for enterprises to get comfortable, with this new technology. You know, I go back to like, Cloud. AWS came out in 2006 and, you know, didn't really see, I would say, mainstream adoption until like 2015.
So, it took almost ten years for Cloud. I think this will be quicker, but we're about three years into it right now and, into AI, and, you know, so let's say it's seven years and not ten years. We’ve still got another, you know, few years, I think, before we see, you know, broad adoption in the enterprise of AI, AI apps, and AI tools. The biggest, sort of, adopt adoption area right now has been coding.
So that's, sort of, the tip of the spear. And that's being very widely used, and certainly, driving a lot of efficiency and excitement. But I would say broader adoption is still, I'd say, you know, a couple of years away.
18:23, Richard
Yeah. I mean, I was having some meetings with some clients, and one of them said, I mean, ultimately it could be slow because the company, you know, if you suddenly switch to, you know, an AI provider as opposed to your software provider, you no longer have a throat to choke. I don't know if you've heard that one, but, you know, so that that, kind of, you ultimately need someone to blame. And if the cost is not there and if it's, you know, fits in with compliance and regulatory. So, I think you're right, that's going to be slow.
So, where do you think we are now on valuations for these for these stocks? Do you think we're, sort of, nearing a bottom now and or, you know, much more attractive?
19:05, Jason
I mean, that's always a hard, you know, that's a hard one. It's, sort of, call the bottom. So, I don't want to do that. What I will say is certainly the multiples have rerated pretty substantially, where, you know, like, let's call it average software company might have been something like six times revenue, and now it's like three to four times revenue, so pretty substantial rerating. I think it might take, you know, sort of, the M&A engine to, kind of, get going to help, sort of, figure out where the bottom is, right? in terms of what people are willing to pay, either strategic or private equity. I think private equity right now has its own issues. So, we might see a little bit less, sort of, private equity activity in terms of picking up some of these, let's call it depressed software assets on the public side.
But, historically, you saw that that was a, sort of, a backstop for software company valuations. Private equity would come in and, you know, pick up some of these companies. And I think that's another fear, by the way, which is, like, the private debt markets, which, you know, there's, like, a heavy concentration of software in a lot of that private debt market.
And so, I don't want to say we've hit the bottom. But if the numbers, you know, develop in the way that we think they can this year, and companies show upside to what they've guided to, and certainly in the companies that can accelerate revenue growth, that's going to help, you know, investor confidence on, sort of, the durability and viability of software sector.
But it is, you know, I don't know that we've seen, like, capitulation. But, you know, there's been a lot of rotation out of software for sure.
20:45, Richard
And I mean, have companies changed their guidance much? Have you heard much difference there?
20:52, Jason
I mean, it's case by case, but, you know, and sometimes you have some idiosyncratic things that have nothing to do with AI. People, you know, kind of, fear the worst. But, in general, like my companies this quarter have beat numbers and, you know, maintained or raised guidance. So, you know, it's still feels pretty good out there, despite, you know, some of these, kind of, amorphous fears.
21:19, Richard
Thank you very much, Jason. That was super helpful, super insightful, helped calm my nerves a little bit.
21:33, Chris
Jason, thanks so much for joining. Richard, as always, good to have you. Thanks to everybody who's been listening. We'll be back next month with another episode. Take care.