SaaS Is In Even More Trouble Than The Hype Would Have You Believe

The current debate about whether “SaaS is dead” has two camps. Camp one: SaaS is in serious trouble at the hands of AI. Camp two: SaaS is just fine, nothing to see here, folks.

Both camps are wrong. The truth is more interesting — and more concerning for incumbent SaaS companies — than either side admits.

The Stock Market Is Sending a Signal

Start with what the public markets are telling us. Salesforce is down over 40% from its highs, and the slide has been relentless — dropping 20% in 2025 and another 10%+ to start 2026 as AI monetization concerns shake the software sector. Adobe is in even worse shape —down 34% in the past year, hitting a seven-year low after being branded an “AI loser” by analysts, with its CEO stepping down amid the turmoil. These aren’t small corrections. Something structural is happening.

Meanwhile, I have two charts from Goldman Sachs research worth describing (I’ll link the images below). The first, from Goldman Sachs and Cathay Capital, shows traditional software revenue growing more slowly over the next five years, with most of the growth coming from agentic AI layers. The second, from Goldman Sachs and Gartner, shows profits in the software industry starting to shift from the software layer to the agentic layer over the same period.

So even the optimistic projections say SaaS will still grow — just less so than before, with AI eating an increasing share of the value.

But This Isn’t Just About AI

Here’s where I part ways with the clickbait. This is the same tired “X is dead” framing we see applied to pockets of technology every few years — and it’s always overstated. The truth is that SaaS is under pressure from multiple directions, and AI is only one of them. (For fun, I looked on my blog for articles about Email is Dead and found a TON of them, including this one from – get this – 2004 – “not dead yet,” to quote Monty Python).

Next-gen SaaS companies are winning over legacy incumbents. HubSpot is growing 20-25% year-over-year while Salesforce manages 8-11%. That’s not AI disruption — that’s a nimble competitor beating a slower, older one with more baggage and less innovation. Some big guys are in trouble, but the up-and-comers are thriving. This has been true in every era of enterprise software.

SaaS is overbuilt. As with packaged software in the 2000s — and as I’m sure will happen with AI soon enough — SaaS has become the “land of 1,000 niches.” Sure, you can justify one more $10K enterprise SaaS spend if you narrowly look at the ROI of a single point solution. But eventually, you have a stack that’s overbuilt, crowded, hard to manage, and stuffed with too many overlapping tools. Point solutions eventually get absorbed into the functionality of larger platforms. This is a natural cycle, and it was coming with or without AI.

Data point of one, but our own internal GTM stack at Markup AI is going through a massive overhaul right now. We are replacing 15 vendors with 4 vendors. We are going to reduce our overall SaaS spend by 60-70%. And we are doing it inside of six months with limited business disruption or risk. Yes, it’s a data point of one, and we’re at best a mid-sized buyer. I’m always skeptical about extrapolating from a single example — but in this case, what we’re doing CAN’T possibly be that unique.

Forbes reported recently that Retool found 35% of companies have already replaced at least one SaaS tool with something they built themselves, and 78% plan to replace more this year. That shift was underway before the latest wave of AI developments. The enterprise appetite for bloated SaaS stacks was already waning.

Chris Dunlop made a smart point that the “AI kills SaaS” narrative is lazy. What’s actually happening is more nuanced and multi-causal. I agree — but I’d add that the nuance makes it worse for SaaS, not better. Multiple disruptions hitting simultaneously is harder to survive than a single one.

And Yes, AI Is Going to Make It Worse

Now let’s talk about AI specifically, because it IS also going to disrupt SaaS — significantly.

If you think about it, SaaS is basically complex workflow on top of an amalgamation of data stores. AI is going to disrupt that by allowing people to customize workflows on the fly, as long as systems can access the underlying data. Why pay for a rigid, pre-built workflow when an AI agent can construct the workflow you need in real time?

It’s possible that big SaaS companies will respond by injecting enough AI into their systems to stay alive. Some will. But if they have to do so while facing disruptors who are beating them not just on functionality but on price — and remember, we’re reducing our GTM stack spend by 60-70% — that’s not a small change to the SaaS universe.

It’s not just that the sector will grow more slowly or get less profitable. It’s that parts of the sector are going to shrink at the hands of AI. That’s a fundamentally different story than “SaaS is fine, it’ll just evolve.” Some of it will evolve. Some of it will disappear. But it will get smaller, not just grow less.

The companies pretending the challenges don’t exist are fooling themselves.

P.S. I’m slightly terrified that the Private Credit issues we’re seeing in this country have this as an underlying cause, which if broad enough could make this a canary in the coal mine for a systemic financial crash or correction. But that’s (hopefully not) a subject for another day.