AI's not gonna kill SaaS, but it will force some changes

The following post is a collaboration with my long-time friend, colleague and writing partner Alex Wilhelm. Subscribe to Alex's newsletter ‘Cautious Optimism‘ here. You'll be glad you did!

It's not exactly a secret that SaaS stocks have taken a pounding in recent weeks. Investors worry that as agentic AI (AI-powered software that can autonomously perform tasks) matures, companies will be able to simply create bespoke software themselves. 

But building enterprise-grade software is harder than it looks. You need to worry about security, compliance, identity, observability, integration with other software and more. It's not easy. It's why companies pay big bucks to have others build it for them. 

But two things can be true at once: companies can build their own enterprise software, and it can still be extremely hard to do well.

In many ways, it feels like we’ve seen this movie before. Early in the cloud era, I remember IT pros swearing they’d go to the cloud over their dead bodies. Yet a few years later, most were on their way (and were very much alive). The resistance didn’t stop the shift — it just delayed it.

Look no further than Siebel Systems, the dominant on-prem CRM player in the 1990s, before Salesforce moved the category to the cloud and helped usher in the SaaS era. Just as Salesforce displaced Siebel, it’s possible investor fears are justified and a similar dynamic is forming now.

​​As Tien Tzuo, founder and CEO of Zuora, and Salesforce employee #11, told me, you have to be prepared for a transformational change like AI. He even wonders if entire categories of software could disappear again. “Every software company is facing an existential threat right now, including us, right? And you're better off meeting it and seeing it and navigating through it,” he said.

Agentic AI may indeed be a serious threat at some point, but that doesn’t mean SaaS stocks deserve the beating they’ve taken in recent weeks. So we took a look at five enterprise SaaS stocks — Salesforce, ServiceNow, Adobe, Workday and SAP — and reviewed their numbers over the past three quarters to see what the fundamentals actually look like. Each has embraced AI. They’re clearly not running away from it.

To SaaS or not to SaaS

If SaaS companies are transforming with the times, the real question is: Are investors behaving irrationally, or are these companies under threat and simply getting the value haircut that they deserve?

To understand their respective progress, let’s drill into their most recent quarterly growth rates and highlight critical AI metrics. Then we’ll discuss the three most common concerns about SaaS and how they apply, or not.

• Salesforce: The CRM giant reports Q4 results on February 25, so we’re looking at the first three quarters of its fiscal 2026. Salesforce saw revenue grow 8% in its first fiscal quarter, 10% in its second, and 11% in its third. During those periods, the company also announced that its Data Cloud (now Data 360) and AI products (Agentforce) grew from annual recurring revenue of $1.0 billion in the first quarter to $1.2 billion and $1.4 billion over the second and third quarters.
• ServiceNow: With its full 2025 results in hand, we’re looking at ServiceNow’s second, third, and fourth quarters. ServiceNow revenues grew 22.5% in Q2, 22% in Q3, and 20.5% in Q4. Its ‘Now Assist’ service, which the company describes as an “out-of-the-box GenAI” tool, beat company ACV expectations in the second and third quarters, and finally hit $600 million in the fourth.
• Adobe: Adobe’s revenue grew 11% in Q2, 11% in Q3, and 10% in Q4 of fiscal 2025. The company also reported that “new AI-influenced ARR now exceeds one-third of [its] overall book of business,” after its “AI-first ARR” broke the $250 million barrier in the third quarter, exceeding the company’s own year-end target.
• Workday: Workday’s revenue grew 12.6% in Q1, 12.6% in Q2, and 12.6% in Q3 (how often do we see that level of consistency!). In its Q3 earnings call, Workday said more “than three-quarters of net new deals and 35% of customer expansions included one or more AI products,” and that more than three-quarters of its “core customers are using Workday Illuminate AI, driving well over 1 billion AI actions on the Workday platform this year alone.”
• SAP: Finally, SAP’s revenue grew 9% in Q2, 7% in Q3, and 8% in Q4. (growth rates were 12%, 11%, and 11%, not accounting for currency fluctuations). 

The major SaaS players that we’ve touched on, except Salesforce, are not seeing revenue accelerate in the AI era. Yet. What we are seeing is ~static revenue growth rates as AI revenues come online and scale. From that perspective, the SaaS world is showing that it can maintain revenue growth with new AI incomes.

That’s why the idea that SaaS is about to dry up and blow away is incorrect. If vibe-coding were primed to, or already starting to, eat the best-known, traditional SaaS products (CRM, ERP, HR, etc.) we’d have likely seen the negative effects on revenue by now. 

Yet, ServiceNow beat its own Q4 growth expectations, Adobe raised its growth target in the second quarter, Workday raised its fiscal year guidance in its second quarter, while SAP expects revenue to rise faster in its new year.

Stop making sense

How can we make sense of declining SaaS share prices if the underlying assets aren’t falling over and fainting in the face of new competition empowered by AI? Investors may simply be more worried than these results warrant, but several confluent factors are worth considering:

1. AI coding tools are rapidly improving. The theoretical risk of current and future SaaS customers building their own replacement tooling is increasing.
2. Companies want fewer humans: The reticence to add headcount amongst major companies today means fewer new added seats. Corporates getting excited about cutting staff could mean more seat churn.

What’s interesting about those two concerns is that they do not undercut our core thesis: Companies can build their own software, but it’s hard to do so.

Why is it hard to do? You may be able to abstract away some of the complexity of building and maintaining in-house code as AI improves. But you can’t vibe code the data and many companies still struggle mightily when it comes to wrangling data . That’s why SAP CEO Christian Klein took the time to discuss where his company sees strength in the AI era instead of weakness (lightly edited and condensed for readability):

Now, talking about the future of SAP AI, talking about the future of SAP, and I know there is a general concern out there in the market about how will software sustain in the world of AI? Cannot everyone code software? I would say clearly, no. Because what we are already seeing with many customers is [they are] building certain custom agents for cash flow collection, et cetera, with those LLM providers. But what you always see as a roadblock [...] What [about] sales negotiations, deals in the pipeline? What about certain payment information, which are also necessary for the agent to understand why is this customer not paying? So it always goes together. The LLMs are super good [with] unstructured data, but you need the business data. And which company has petabytes of data, which we are using to fine-tune our AI Foundation?

Not only are traditional SaaS companies sitting atop an ocean of data (which may explain why Salesforce is accelerating a little), they have a leg up on building AI-first software. Sure, you could vibe code an SAP replacement, but it won’t have your data, and your LLM won’t have the same data history as the incumbent. Naturally, the large SaaS shops could blow this advantage, but parsing their statements, it seems each company is taking this leg up seriously.

Sure, you could extricate your company’s data from a vendor and run your own version of their software in-house. It might even work. But you’d be forced to leave on the table what Klein said at the end of his riff: the value of the centralized SaaS company building AI tooling on top of aggregated customer data, not just the information of a single entity . 

So even if you get all your data back and build your own cheaper tool, you’ll have to admit that you won’t be able to learn from what other companies are up to. SAP can, and it’s betting it can do better with an ocean of data than any customer could with a pond filled with only their information.

None of that may be enough to sway investor sentiment in the near term. But merely seeing SaaS companies defend growth rates and talk about reacceleration is telling that the sector is not headed for the wood chipper. Perhaps investors need to look beyond the hype and lighten up.

Yet, in spite of the obvious data advantage that SaaS companies have, the coding agents really are getting good. If AI developers can begin to solve those very real last mile problems, perhaps SaaS companies will have something to worry about. For now, though, Wall Street could be putting the cart well before the horse.