Overview
"90% of SaaS will disappear" sounds like doom-bait, but the logic is sound: a large share of today's SaaS is undifferentiated, and several forces are converging to wipe out the weak. This report explains the shakeout and what survives it.
AI floods the market
When building software gets dramatically cheaper, far more of it gets built. Any feature that's easy to clone with AI becomes a commodity overnight. Thin "wrapper" products — a prompt and a UI on top of a model anyone can call — have no defensibility and will be the first to die.
Platforms absorb features
The "sherlocking" pattern is brutal: when a single-feature tool gets popular, the platform it sits on (the OS, the suite, the model provider) builds the feature natively and bundles it for free. Single-feature SaaS lives at the mercy of the platform beneath it.
Buyers consolidate
After years of tool sprawl, companies are cutting their stacks to save money and reduce complexity. Procurement consolidates around fewer, broader vendors. The marginal point tool with overlapping functionality gets cancelled at renewal.
What survives
The survivors share defensibility: proprietary data that improves with use, deep workflow integration that's painful to rip out, distribution advantages, or network effects. They solve a real, painful problem and own something competitors can't cheaply copy.
What this means for you
If you build SaaS: assume your features will be cloned and ask "what do I own that they can't?" Invest in data flywheels, integration depth, and distribution. If you buy SaaS: expect churn among your vendors and favor ones with durable moats. If you invest: discount thin wrappers heavily.
Honest limits
"90%" is directional, not precise — plenty of niche, sticky SaaS will thrive. The point isn't that SaaS is dead; it's that undifferentiated SaaS is, and the bar for durability just rose sharply.
