A software category is about to be wiped out. Not the horizontal platforms. Not the infrastructure. Not the CRMs and ERPs that have become as embedded as electricity.
I'm talking about the $20M ARR sweet spot — the tidy, vertical SaaS product that built a comfortable business solving one specific problem. Field service management. Contract lifecycle. Sales territory planning. Compliance tracking. Employee onboarding.
They grew to a respectable scale by solving real problems at a moment when building custom software was genuinely hard. That moment is ending.
The Math That Made Buying Obvious
For twenty years, the build-vs-buy calculation was simple: building was slow, expensive, and risky. A $20M ARR vendor had forty engineers who had already solved the problem you needed solved. Buying gave you a working product on day ninety. Building gave you a scope creep nightmare and a half-finished system eighteen months later.
AI has broken that math.
A competent team using modern AI-assisted development tools can now produce production-quality, well-architected software at a speed that was unimaginable three years ago.
What You're Actually Paying For
When you buy a $20M ARR SaaS product today, you're not just buying software. You're buying a lock-in you haven't priced yet. Every integration your team builds is a switching cost. Every workaround you create because the platform won't do exactly what you need is a tax on your operations.
And the pricing is structurally adversarial. As your usage grows, so does your bill.
The Disruption Pattern
Spreadsheets killed a generation of specialized reporting tools. Cloud killed on-premise software. The pattern is consistent: the companies that get disrupted first are the ones that solved a real but narrow problem at a moment when that solution required significant capital and expertise to build. When the barrier drops, the business model collapses.
The signal is already visible inside companies. When your team starts asking "could we just build this?" — and the honest answer is increasingly "yes" — you're looking at a category that's about to tip.