“Reports of My Death Are Greatly Exaggerated”

By Mark Strauch, Founding Partner and President, Alpine Investors

Everyone knows AI will change the software industry in profound ways. It’s already happening. Just look at the dramatic signal from the capital markets: SaaS indices have traded down 25% since September 2025.¹ The so-called AI “SaaSpocalypse.” But it’s time to take a closer look, because SaaS is not—in fact—dead. Rumors of its demise are greatly exaggerated.

One interesting data point was the 2025 surge in SaaS M&A volume. Per Pitchbook—and cited in Allie Garfinkle’s recent piece on the “death of SaaS”—enterprise SaaS M&A hit $83.7 billion in total value in Q4 2025, with a nearly 24% leap in quarter over quarter deal value. This made 2025 the most active year for enterprise SaaS M&A since 2021.²  Now, as we look forward to 2026, buyers may view SaaS as a growing opportunity to buy at attractive valuations. For the first time in a long time, SaaS multiples are now trading below the rest of the S&P.³

Another interesting datapoint is that software companies continue to perform well. Public company valuations have separated from underlying fundamentals.⁴ Rule of 40 performance has soared while multiples have gone the other direction. In fairness, the SaaSpocalypse narrative was never really about this quarter’s financial performance. It’s more of a fear-based, speculative sentiment based on what may happen in the future. A future expectation that SaaS customers will churn or reprice because the LLMs or the customer will build agents to replace the work that software performs today. A view that SaaS growth rates will slow or reverse because highly recurring software revenues are no longer the attractive, predictable annuity they once were. And as a result, long-term earnings forecasts are now highly questionable and should be discounted heavily. Or so goes the story. To be clear, in my estimation that argument is not wrong—it will play out that way—for some. But the important thing to understand is that it won’t play out that way for all.

The sweeping, across-the-board down-trade is indiscriminate. And that’s the problem. The truth is, there will be winners and there will be losers. The trick is knowing which is which. We can talk at an abstract, high level about things like “rents” and “supplier power” but we also have to fly way lower, and examine what we know about how and why actual customers have chosen to use the software they use. What actual business problem are they solving, what do these customers care about, and how is that changing now in light of AI alternatives? What economic moats are in play and which moats are truly durable going forward? And so it will be the answers that investors, business owners, and SaaS operators offer to these questions which will inform the inevitable “flight to quality” that starts now. A flight that always follows a dramatic risk re-pricing moment like this one. So, what then is “quality?” How do we know where to fly?

Let’s start with who you don’t want to be right now if you’re a SaaS business. You don’t want to be priced by seat. Or serving a broad and generic record-keeping function. You don’t want to be serving enterprise customers who possess large, sophisticated technology capabilities in-house. You don’t want to be a point solution holding your breath hoping that Anthropic’s next plug-in doesn’t land in your backyard. In short, you don’t want to be a seat-based horizontal enterprise software tool.

OK, so what do you want? You want to be deeply embedded in complex business workflows tied to proprietary data and subject to regulatory compliance factors. You want to be highly tailored to the needs of a specific set of vertical market customers who rely on you because they lack the capacity to do it themselves. In short, you want to be a mission-critical middle market vertical SaaS business.

I wrote about this dynamic back in 2020. The core argument at the time was that the most interesting software investment opportunities weren’t the horizontal giants. They were a new generation of “SaaS 2.0” businesses which were vertical, profitable, bootstrapped businesses serving the specific needs of dentists, mental health agencies, or vet practices. Software built for a narrow set of customers with invariably similar requirements, generating upwards of 95% gross retention rates, embedded so deeply into daily workflow that switching would destroy value and be highly disruptive.

At the time, our investment thesis was that these businesses were overlooked and undervalued. Today, I’d make a different argument: those same defining characteristics that made SaaS 2.0 an attractive investment opportunity are exactly what makes these stickier vertical SaaS businesses more defensible against AI disruption in today’s world. Deep vertical specificity, proprietary data built up over years of customer transactions, regulatory complexity, and workflow embeddedness are the competitive moats in an AI world.

But that’s not the end of the story. Let’s say you are an investor or an employee or a leader of one of these mission-critical vertical SaaS businesses. You know the qualities I describe are necessary for survival in an AI world. But while they are necessary, they’re not sufficient. Sufficiency comes when you add one more thing. You want to be incorporating AI into your product so that AI enhances your position rather than threatening to replace it. In fact, when taken to the fullest, you effectively become vertical AI, not vertical SaaS. And when you realize that opportunity, you can deliver important business outcomes and price accordingly. In order to fully deliver on outcomes, you can’t be a point solution, a small piece of a larger puzzle. You need to be the full puzzle for a given customer or a given business process.

So, taken together, the future winners in SaaS will do three things well:

1. Defend their distribution incumbency with the mission-critical attributes of proprietary data, deeply embedded workflow, and vertical domain specificity
2. Advance their product with embedded AI that delivers business outcomes by automating and orchestrating work that is performed by both humans and agents
3. Price their product based on outcomes achieved to unlock more value for the customer and also more growth opportunities for the company

We all know customer expectations will rise. We all know LLMs will continue to release eye-popping advancements: from training to inference to specialized plug-ins, from chatbots to individual agents to agent-to-agent cooperation. And we see private equity firms scrambling for an advantage by announcing ambiguous AI partnerships.⁵

What remains to be seen is which SaaS businesses will endure, flourish, and grow. Which businesses have real and lasting moats? The winners will be those who got ruthlessly honest fast enough to disrupt themselves before someone else does it for them. Warren Buffett’s famous line seems to apply: When the tide goes out, we find out who’s been swimming naked.

—