The SaaSpocalypse: A Misdiagnosed Crisis

The recent market shock known as the “SaaSpocalypse” has left many in the tech industry scrambling to make sense of it all. Beneath the drama and hyperbole, a more nuanced reality is emerging: enterprise software value is not being destroyed by AI; rather, it’s being re-priced.

For years, I’ve watched companies build complex systems that promise to reshape professional work. The SaaSpocalypse debate has focused on the wrong layer of the software stack. A clear distinction exists between where value compounds and where it erodes – if you look at the system as a whole.

A Layer of Irrelevance

Much of enterprise software lacks durable differentiation; it’s just a wrapper. When AI agents can perform tasks that an application was built to support, that application becomes optional. The interface is no longer a screen but a conversation, and multiple models can deliver that interaction. This is particularly true for the application layer, which manages how users interact with information, not the information itself.

The companies being repriced are not being punished due to panic; they’re being re-priced because their differentiation was always in the interface: the interaction model, workflow sequence, or UI skin on top of someone else’s data. When the interface is no longer a screen but a conversation, and multiple models can deliver that interaction, the wrapper carries far less value.

What the Debate Gets Wrong

The SaaSpocalypse narrative stops at the collapse of the wrapper, as if that’s the whole story. But it’s not. AI is actually shifting value downward through the stack, rather than destroying it. The application layer is commoditizing; the intelligence layer is not.

In fact, modern AI systems’ agentic capabilities support a thesis that differentiation in the intelligence layer is growing and durable. As work agents are asked to do more complex tasks, it becomes harder to fake underlying competence. And as agents take on more complex, multi-step work, the gap between intelligence that can be trusted and intelligence that merely sounds convincing widens.

Two Layers That Matter

The two layers that actually matter are the trustworthiness of the underlying intelligence and the knowledge content grounding the work. In professional contexts, accuracy is not a nicety; it’s a requirement. A hallucinated case citation or an incorrect tax precedent can be catastrophic – and general-purpose AI can generate plausible-sounding output.

Domain-specific training, grounding, and validation are essential for trustworthy intelligence. When AI agents need to produce outputs that professionals can verify, the value of the underlying knowledge infrastructure becomes clear. The organizations that have spent decades assembling and maintaining this knowledge are not facing a SaaSpocalypse; they’re looking at a world where their content becomes more valuable.

A New Era of Value Creation

Executives should be asking how to leverage the growing importance of trustworthy intelligence and authoritative domain content, rather than which applications to keep. The organizations that can provide this infrastructure will create value in the long term – not just those who can replicate a wrapper.

The SaaSpocalypse has exposed where software value really lives – but only if we look beyond the drama and hyperbole to the underlying realities of the system.