The SaaS-pocalypse Is Here. Or Is It?
Marc Andreessen said software would eat the world back in 2011, and for a while, he was absolutely right. SaaS companies minted billions, valuations soared, and every business process sprouted a subscription. Then AI showed up and apparently decided it was hungry too. Salesforce is down 21% this year. ServiceNow, 26%. Adobe, 22%. Intuit, a brutal 37%. Investors are doing what investors do when they get nervous: selling first and asking questions later. The question everyone is asking is whether this is a market overreaction or the beginning of something genuinely structural.
The bear case is not subtle. If AI can automate the tasks that SaaS tools were built to handle, why would enterprises keep paying for software licenses? Mistral's CEO has suggested more than half of current enterprise software could be replaced by AI. That's the kind of statement that sends portfolio managers reaching for the sell button. And unlike previous tech selloffs driven by overvaluation or hype, analysts are noting that this one is different because there are actual existential question marks around the underlying business model, not just stretched price-to-earnings ratios.
Here's where it gets more nuanced. The SaaS products most at risk are the horizontal point solutions: tools that do one relatively generic thing at scale. The ones likely to survive are those with deep vertical integration into complex industries like healthcare or manufacturing, or those sitting on proprietary data that AI can't simply replicate from scratch. The idea that a company could rebuild decades of specialized enterprise software in-house using AI tools has been called not viable by more than one analyst. Enterprise software is complicated precisely because the problems it solves are complicated.
From an IT leadership perspective, this moment has real practical implications. Organizations evaluating their SaaS portfolios should be asking hard questions about which tools are genuinely differentiated and which are subscriptions to functionality that AI can now approximate. That audit is worth doing regardless of what happens to stock prices. At the same time, the assumption that AI agents will simply replace SaaS wholesale underestimates the complexity of change management, data governance, and compliance requirements that make enterprise software sticky in ways that don't show up in a demo.
The honest answer is that both sides are probably partially right. Some SaaS products will be disrupted. Others will integrate AI capabilities and emerge stronger. The vendors who treat this moment as a fire drill rather than an extinction event will likely find a path forward. For IT leaders managing SaaS portfolios worth millions in annual spend, now is exactly the right time to be asking which subscriptions are genuinely earning their seat at the table, and which ones have been coasting on inertia and a good renewal conversation.
https://www.cnbc.com/2026/02/27/tech-download-existential-software-crisis.html
