If you’ve spent any time on X or scanning Bloomberg lately, you’ve likely seen the viral take currently making the rounds: “SaaS is dead.” While the sentiment has been amplified by "AI-bros" and Wall Street analysts, the spark arguably came over a year ago from Microsoft CEO Satya Nadella when he said the following in an interview with Bill Gurley & Brad Gerstner:
“…the notion that [SaaS] business applications exist, that’s probably where they’ll all collapse, right in the Agent era…”
Nadella is suggesting that the next generation of AI won't just sit on top of software—it will replace the very concept of "apps" as we know them. The argument goes: Why would a company pay Salesforce or Zendesk hundreds of dollars per seat when a custom AI agent, spun up for pennies, can do the job better?
The idea that custom AI code will replace SaaS platforms at scale is an AI-bro fever dream. I know exactly zero CIOs considering replacing their entire SaaS tech stack with bespoke AI software. Why? Because the ability to replace SaaS with less expensive custom solutions has been available for decades. Choosing to do so means moving from the KNOWN to the UNKNOWN.
Most CIOs are wildly allergic to the unknown. They are the ones who have to answer the pager at 3:00 AM when a server tips over. They aren't going to trade a vetted, supported platform for a "hallucination-prone" custom build just to save money on seat licenses.
For this reason alone, we are many decades away from anything like this happening—but there is a vulnerability in SaaS. It's not because AI is going to rewrite everything; it's something more fundamental. Let’s look at the actual mechanisms at play.
First, let’s acknowledge the elephant in the room: Tech has had a massive run. We are seeing average P/E ratios around x30, which is roughly 15% to 35% higher than the 10-year average. While we aren’t at the "dotcom bubble" x50 levels, seeing NVIDIA at x39 trailing earnings tells us one thing: Tech is due for a correction.
The Real Threat: Recession and Layoffs If the "AI-apocalypse" isn't going to kill SaaS revenue, what will? Two very old-school mechanisms: Recession and Layoffs. We are currently seeing a collision of AI-hype and a largely undisclosed economic slowdown.
When the economy slows, companies look for easy places to save. They don't build custom AI empires; they play a game I call "Software Whack-a-Mole." They reduce headcount and cut unused SaaS seats.
For SaaS companies, this results in a sharply reduced seat count and a direct hit to revenue. This mirrors the survival playbook we discussed in our Google Zero strategy—shifting focus toward direct value and owned channels rather than hoping for growth in a shrinking environment.
The threat to SaaS isn't that AI is "smarter" than SaaS—it's that the economy is cooling, and fewer humans in chairs means fewer licenses on the ledger. As we move further into 2026, don't buy the hype that AI is eating the world, although it may have lit the spark of the economic slowdown that's coming.