Is SaaS dead? The honest answer
The old kind is dying — thin tools, generic wrappers, feature-compete pricing. What replaces it is already visible.
The case for the prosecution
The 'SaaS is dead' argument has three legs, and all of them are real. First, vibe-coding: when a non-technical user can describe a tool and get a working version, the addressable market for simple paid tools shrinks. Second, platform risk: OpenAI, Anthropic, and the model labs keep shipping features that eat the application layer — every capability launch is a small extinction event for the tools built on top of the last one. Third, agents: if software is increasingly operated by other software, the moat of a polished UI evaporates.
Add the economics: SaaS multiples compressed because the market repriced how durable a subscription to a clonable feature-set really is. The 'just add AI to a CRUD app' era ended the moment anyone could generate the CRUD app.
The case for the defense
And yet — somebody still has to run the business. SaaS was never really about the code; it was about outsourcing a problem to people who keep it solved. Support, compliance, reliability at 3am, the integrations that hold a workflow together — none of that gets vibe-coded. A clone has your features; it doesn't have your uptime record, your security review, or the trust of the team that adopted you.
There's also a real counter-current most takes ignore: security and data concerns are pushing serious buyers toward dedicated vendors and even on-prem deployments, not toward 'let the general model handle it.' When the agent era actually arrives, enterprises will want vendors that take responsibility — liability doesn't vibe-code either.
Key takeaway: SaaS isn't dead. The lazy version of SaaS is dead — the version where the product was the moat.
What's actually dying vs what's surviving
The cleanest way to read the debate is to sort the industry by what's underneath the product.
Dying — thin horizontal tools
Generic schedulers, note-takers, form builders, dashboards. A weekend clone or a model update replaces them.
Dying — AI wrappers on public data
Nothing underneath the API call. The model's next release is the competitor.
Surviving — vertical SaaS with proprietary data
Deep in one industry's workflow, holding data that only exists because the product is used. High retention because ripping it out means losing the data.
Surviving — distribution-first products
Average product, excellent reach. Beats excellent product, no reach — every week, in every niche.
Surviving — anything with real-world coupling
Payments, compliance, physical ops, community. Bounded by atoms or trust, not code.
What it means if you're building now
The 'is SaaS dead' question resolves to a simpler one: what's under your product? If the answer is features, you're building on melting ice — enjoy the revenue while it lasts and don't confuse it for a company. If the answer is proprietary data, real distribution, deep workflow embedding, or a community, SaaS isn't dead for you at all — the shakeout is actively removing your competition.
The meta-lesson of the debate: every founder now has to answer 'what's my moat' on day one, because the market finally stopped giving credit for effort. The good news for small teams is that the surviving moats — data, distribution, niche depth — are the ones money can't shortcut.
Key takeaway: Don't ask 'is SaaS dead.' Ask 'is my SaaS a moat or a feature set.' The market is already asking it for you.
Build the part that survives
Distribution is the moat you can start today. List free and put your product where the buyers already look.
Related articles
Building something? Tell the story.
Write about what you're shipping — posts are indexed, linked to your product, and read by people looking for new tools.
Write a post