Why Great Software Stocks Suddenly Stopped Working

0:00 Now, do a little thought experiment with me here. Imagine you're looking at a stock. You're monitoring the underlying business. Over the past 12 months, they did roughly $9 billion in revenue, a gross profit of 7.7 billion, and that's an 85% [music] gross margin, and an operating income of roughly $2.8 billion, sitting at roughly 31% margins. Pretty solid numbers, right? Now, fast forward 7 years ahead, revenue has just crossed 23.7 billion, more than 2.6 x what it was 7 years ago. Gross margins have actually improved from 85 to 88%, and operating income is up to 8.7 billion, and margins right now are sitting at 36%. So, the business, by every measurable metric, has gotten significantly better. And here's my question. What do you think happened to the share price over the same 7-year time period? So, I'll give you 2 seconds. Drop your guesses in the comments below.

1:04 Done? Now, here's the million-dollar reveal. It went nowhere. Zero. So, a business that more than doubled its revenue, expanded its margins, and compounded its profits, and the stock has essentially nothing to show for. And you might think that this is a thought experiment, but it's a real-life case study. The stock in question is none other than Adobe, the platform that help you look skinnier in photos after a bit of post-production magic. So, I heard that the like button is pretty magical, too. So, you can tap it lightly to see what happens to it after.

1:48 For those of you who are in tune with market developments, this is not an Adobe-only story. So, some of the most recognized and successful software companies on this planet Earth are experiencing a very similar fate. And some investors might even call this the lost decade. So, the business has grown, revenues are up, earnings report continue to looks resilient. So, you did your homework and by every fundamental measure, the business you bought is a better business today than the day you purchased it. And yet, the stock doesn't reflect the same reality. So, what on Earth is going on? So, to understand why investors are so rattled right now, we first need to go back to basics. Because before all of this has happened, software stocks in general are some of the most loved businesses. And for good reasons, of course. Let's start with the business model itself. And by the way, if you enjoy this kind of breakdown, we have a weekly newsletter called Investing Bytes, where we do exactly this. Using real-world examples, providing you with clear thinking frameworks. So, the link to sign up is in the description below and it's completely free. Now, let's get back to the video. So, most traditional businesses have what you call a one-and-done relationship with their customers. So, if you think about a car manufacturer, they build a car, they sell it to you, and that's largely the end of the transaction itself. So, in order for them to grow, they probably need to find new customers, build more cars, spend more money on building new factories, and hire even more workers.

3:19 So, every unit of growth actually cost them additional incremental amounts of money. Software, however, on the other hand, doesn't work this way at all. So, think about what a company like Salesforce actually sells. It's code. Lines and lines of code. And here's the beautiful thing about coding. Once it's written, it costs virtually nothing to deliver to the next customer. So, the 10,000 customer costs almost the same to serve as the first one you get in the door. And you essentially just send them a login, and there you have it, the transaction is done. That's why software companies routinely report gross margins of 75, 80, or even 85% and above. And for Adobe's case, we just saw they're sitting at an 88% gross margin. And that's not normal, by the way. That's extraordinary relative to other companies, of course. But the margins are only half the story. The other thing that actually makes software companies so special was how these companies charge their customers. So, rather than a one-time sale that we've seen in the car manufacturer perspective, almost all of these software companies moved to a subscription model. So, you pay monthly or annually, and as long as you keep using the product, you keep paying. And here's the kicker. People almost never leave. And ServiceNow, for example, renews roughly 98% of their customers every single year. 98%. You didn't hear that wrongly. I don't really think you can really consider this situation customer loyalty. That's practically a hostage situation. And this dynamic in the software space has created something investors absolutely dream about: predictable and recurring revenue. And a software-as-a-service company, SaaS in short, could enter January 1st already knowing, with reasonable confidence, what a majority of their revenue for the entire year would look like before making any new sale. So, for a business analyst or a fund manager trying to model out the future cash flow of this company, that kind of visibility is seriously some sort of a gift from above. And then, there's one more additional layer that really supercharged the growth story. And it's an industry term used, and it's called net revenue retention. And the concept is relatively simple. So, over time, existing customers of the software businesses don't just stay. They tend to actually increase their spend as they get comfortable with the platform and unlock new features, and then thereafter upgrade their plan. So, even if it's a software company that signed zero new customers in any given year, they'll probably still grow their revenue just from their existing customer base expanding their usage. And Snowflake at its peak was reporting a net revenue retention of above 150% meaning to say every existing customer is spending 50% more year-over-year all on their own.

6:09 And if you were to put all of this together, extraordinary margins, predictable recurring revenue, customers that barely leave, and an existing base that grows itself. You'd pay a premium for that kind of companies, too, and investors did. And these companies routinely trade at at roughly 15, 20, sometimes even 30 times their annual revenue multiple, revenue by the way. And by normal standards, that sounds pretty absurd. But a model this clean, this predictable, this capital-efficient, I think it's quite fair to understand why the market value it this way. And that's almost where the story started to take a turn. For most of 2024 and in 2025, the AI conversation was mostly background noise for software investors. Yes, we saw ChatGPT being birthed out in 2022. Yes, everyone was talking about it, but the general consensus was that these big software companies will just include AI features into their existing suite of products, charge a little more for it, and there you have it, life goes on. So, the business model itself survives and maybe it even gets better, who knows? But something really started to shift in early 2026, Anthropic, one of the leading AI companies in the world, released a suite of autonomous AI agents capable of handling complex multi-step business tasks entirely on its own, and that's why it's called autonomous. So, we are talking about things like lead generation, contract review, legal auditing, so on and so forth. And the legal software sector got hit first, and it got hit really, really hard. So, London Stock Exchange Group, which operates a large data analytics business, fell 13% in a single day.

7:53 Thomson Reuters dropped 16%, and you have other smaller legal tech names fell by 20% or more. And the market started asking a question it had never seriously asked before. If AI can do the work, do you really still need the human doing it? And if you don't need the human, do you still need the software subscription that was initially attached to that human? Now, at this point in time, a lot of the conversation in the media and on social media platforms started gravitating towards something called vibe coding. The idea being that because anyone can now build a software product by typing a few prompts into an AI tool, these big AI software companies would suddenly face an overwhelming wave of new competition. And basically, the lines of code will start commoditized overnight. Honestly, I think that's a bit of an oversimplification. Here's why. What makes Salesforce a $200 billion company was never the code itself. It was decades of customer relationship, enterprise sales team, the integrations, the institutional trust, the weight of the data the customers had built inside the platform over years.

9:01 You probably cannot vibe code your way to that. And the fact of the matter is that there always has been fierce competition in the software industry. And the strongest company survive not because of their code, but their entire ecosystem. But here's the thing. While everyone was busy debating vibe coding, there are three very real risks that the market is attempting to price in today, and in a very, very aggressive sense. Just take a look at the group of software companies and how they've performed thus far. And this is exactly what has caused more software investors to panic and the stocks to bleed out the way that they had. Let's go through them one by one. Fear number one, the death of the seat.

9:40 Remember how we talked about the subscription model initially? How software companies actually charge per user, per login, or per seat? A company with, say, for example, 500 employees using Salesforce pays for 500 Salesforce seat. Simple. Now, here's the problem. If AI agents can do the work of humans, companies need fewer humans. And if they need fewer humans, they need fewer seats. And it's not that AI replaces the software entirely. It's that AI reduces the headcount that was attached to the software itself. And intuitively, when the headcount starts to shrink, so does the revenue of these software companies as they sell per seat. And this isn't a hypothetical assumption anymore. We're starting to see some mid-market software companies recently reporting their enterprise seat count declining for the first time in their history. And you probably would have guessed the stock got absolutely hammered for it. And think about the broader implication here. So, if you're a Fortune 500 company and your marketing team of 100 people can now do the same output with just 60 people because AI is handling a significant chunk of the workload, that's a 40% revenue haircut for Adobe with no change in how good or bad their product is. And this is what keeps SaaS investors up at night right now. Fear number two, the budget rotation. So, for the better part of the last two decades, enterprise software spending operated on a pretty simple assumption. If your team needed a tool to do their job, you went out into the market and you bought the best software available for that job. It was a must buy. The market was there, the vendors were ready, the entire system and pipeline is already done for you. So, with AI dramatically lowering the cost and complexity of building softwares, companies are now genuinely reconsidering what they actually need to buy versus what they could reasonably build for themselves. Not for everything, but for enough things to matter. And this creates a dynamic that is new for software vendors. So, instead of sitting comfortably as the obvious answer, they now have to fight for their position in the budget conversation in every company. So, every renewal is a new negotiation, every contract is being scrutinized, and the question on every CFO's desk is no longer just which vendor do we use, but do we actually need a vendor for this at all? And now, for the big, deeply embedded platforms, like your Salesforce, ServiceNow, Adobe, this pressure is probably relatively manageable because the timeline wouldn't be so jam-packed, the switching costs are high enough, and the complexities are deep enough that building in-house remains an unrealistic alternative. For now, they're still the default answer.

12:23 But for the layer of software companies sitting just below that, the ones that solve problems that are real but not mission-critical enough, the ones with thinner switching costs and simpler workflows, I think this is really an existential question for them. Because the moment a company decides to build something good enough internally, that vendor doesn't just lose a renewal, they lose it permanently, and they never even see it, and they probably can't even see it coming through a churn report.

12:49 Because the customer didn't switch to a competitor, they just quietly disappear. Now, software companies are clearly playing defense in a game that they've spent 20 years playing offensively, and the market is pricing that discomfort in. Moving on to the last fear, fear number three, the margin problem. So, we established earlier that software companies have beautiful margins, sometimes 80, 85, even 88% gross margins. Because the cost of delivering software to each additional customer is basically zero. And here's the uncomfortable reality in this entire AI conversation and even transition. So, to stay competitive, every major software company is now integrating AI features directly into their products. So, Adobe adds generative imaging tool. Salesforce builds AI agents into their CRM. And ServiceNow actually automates workflow with AI as well. And their customers, by default, expects them to do it. But those AI features are not free to run.

13:48 Every time a user generates an image in Adobe, an AI agent completes a task in Salesforce, that requires compute power. And that compute has the relevant cost as well. And Adobe, in this case, has already started moving towards what we call a generative credit model, where customers pay per AI output, rather than just a flat subscription fee. It's usage-based, it's consumption-based, whatever you want to call it. And the point is that a portion of that every new incremental dollar earned from those features actually gets paid out to a third party running the model itself.

14:21 So, the margin profile that investors used to fall in love with, it gets a little murkier and raises a genuine question in terms of the earnings power and the associated margins. And then, to add another layer on top of this, if the same AI feature makes each employee significantly more productive, and we already talked about how it reduces seat count, then software company faces a second pressure, or you can call it a double whammy. You have fewer seats on one side, and you have lower margins per seat on the other. And if that combination actually plays out, it's a meaningful structural headwind for the entire industry. Now, to be fair, these companies are not sitting on their laurels thinking that they wouldn't be disrupted. Many of them are actively trying to reprice their products around outcomes and efficiency, rather than just the traditional head count. And the argument being, if our software makes your team twice as productive, we should be able to charge more per seat, not less. And that argument probably has certain merits. But clearly, in the stock market today, the investors are not willing to give these companies the benefit of the doubt in this entire AI transition. And it wants to see proof first. And until that proof actually arrives, it is pricing in all these associated risks pretty aggressively.

15:37 So, after all of this, I think the question that you're probably sitting with right now is whether this is a buying opportunity for this group of software companies, or is this the beginning of something much more structurally damaging for the entire sector. And my honest answer, it's both, just not for the same companies. Here's the framework we're using to think through this entire problem. There is a dividing line running through the software sector today, and it separates two very different types of businesses.

16:05 So, on one side, you have software companies whose revenues are fundamentally tied to human head count. On the other, you have software companies whose revenues are tied to business activities and state. And that distinction actually matters right now because AI is compressing head count at a very, very quick pace. It is, however, not at least at the same pace of compressing business activity. So, a company might need 20% fewer employees to run their operations in an AI-enabled world, but that does not necessarily mean that they're doing 20% less business. If anything, they might potentially be doing a lot more. And Shopify is a good example of the second type. Their revenue actually scales with the volume of commerce flowing through their platform, not with how many employees they have. Other companies like Adobe, ServiceNow, and Salesforce, they tend to sit in a much more complicated middle ground. So, a part of their legacy or previous business model is head count dependent, while other parts of it scales with activity and outcome. So, their survival story is really much more an execution story instead. Can they successfully reprice the product around outcomes and efficiency before their seat dilution starts to eat into their revenue base? I think the honest answer is that we don't know yet. But these are not small fragile businesses. They have the resources, the customer relationship, and the institutional trust to attempt that transition. Then, there is the other side of the dividing line. The simpler, cleaner, less embedded software businesses. And for those companies, the question being asked right now are legitimate and serious. And I don't think the market is being entirely irrational and discounting them aggressively. Now, here's something worth keeping in mind. The market is a panic machine. When it gets scared, it doesn't reach for a scalpel. It reaches for a sledgehammer. It slashes through everything they see. And right now, it is hitting every software company with the same force regardless of their actual exposure to these associated risks. And this indiscriminate selling is uncomfortable to sit through. But historically, it is also where the most interesting opportunities tend to emerge. And not because the fears are wrong, but because the market's refusal to distinguish between the vulnerable and the resilient creates mispricing that patient informed investors can take advantage of. And the keyword here is being informed. So, just to point out, this is not the moment to buy software stocks because they look cheap on the surface level. Cheap can always get cheaper, especially when the underlying business model is genuinely in transition. And the better question to ask about any company you're looking at right now is where does this company sit on the dividing line? How deeply embedded are they in their customer's operation? How much of their revenue will be affected in this new AI-powered world? And most importantly, are they actively building a credible path to pricing for the future? And if the answers to those questions are reassuring, then yes, the current environment may well represent a genuine long-term opportunity. What is being repriced is the market's confidence in the durability of the model that capture that value. And sometimes when the market starts to lose confidence, that is exactly when the best investments are made. And hopefully, today's video gave you a slightly better map for doing exactly that. And if you want to keep building on the kind of framework we talked about today, we have a weekly newsletter called Investing with Bytes where we break down real world case studies. And if you have stuck around until the end of this video, I have a feeling it's exactly your kind of thing.

19:39 So, hit the link in the description below and we'll send the latest issue straight to your inbox. And this is CK from Piranha Profits signing off. Till next time, keep winning.