SaaS is Becoming Restaurants
Anyone can cook. The recipes are public, the ingredients available at any grocery store. And yet most restaurants fail within their first year. The ones that survive do so through genuine quality, relentless consistency, or pure convenience economics - the pizza place that's just there when you're hungry and don't want to think.
SaaS is becoming restaurants.
The standard narrative about software-as-a-service assumed that building was hard. You needed specialized knowledge, significant capital, months or years of development time. This created natural moats: once you'd built something useful, the difficulty of replication protected you. Customers paid not just for the tool but for the impossibility of making it themselves.
That calculus is shifting. AI agents can scaffold applications in hours. The technical knowledge required to build a functional tool has collapsed. The ingredients are available. The recipes are public.
This doesn't mean SaaS disappears. Restaurants didn't disappear when everyone gained access to the same ingredients chefs use. But the industry changed. Margins compressed. Competition became brutal. Survival required either genuine excellence or operational efficiency so tight that convenience alone justified the price.
What persists when building gets cheap?
Operational burden transfer. The value of a restaurant isn't just the food - it's not having to shop, prep, cook, and clean. SaaS may survive less as "capability you couldn't build" and more as "maintenance you don't want to handle." The question becomes: is this important enough that I need it working reliably, but not core enough that I need to control it?
Network effects. Some tools gain value from other people using them. This is harder to replicate alone, regardless of how easy building has become.
Continuous justification. The era of "build once, extract rent forever" may be closing. What replaces it is "earn it every month or they build their own." Winners embrace this reality and price accordingly. Losers keep acting like building is still hard.
There's a middle zone where SaaS traditionally lived: important enough to need reliability, not critical enough to need control. This middle may be getting squeezed.
If something is truly critical, you need to own it top to bottom. If it's not critical, you might tolerate a self-built solution that goes down sometimes. Where does paid software fit when the space between those poles compresses?
I suspect a lot of existing SaaS was in what we might call the casualty zone - worth paying for when building was hard, not worth it when building is easy. The tools that survive will be the ones that correctly identify which category they're actually in.
A caveat, honestly held: this entire analysis rests on an assumption I can't fully verify. "Agents make everything easy" is a claim that hasn't been stress-tested against eighteen months of actual maintenance.
Building the initial thing and keeping it running are different problems. SaaS bundles both. The people declaring SaaS dead may be overweighting how much of its value was in the building versus the maintaining. When someone's self-built tool breaks at 2am and they realize they're now on-call for their own internal software, the calculation might shift.
Or it might not. The maintenance burden might also get easier. I genuinely don't know.
What I do believe: software is becoming like food. Anyone can make it. Most attempts will fail. The ones that survive will do so through quality, convenience, or accepting margins thin enough that competition becomes the default state rather than the exception.
The SaaS companies that thrive will be the ones that understand they're running restaurants now. The ones that fail will be the ones still acting like they're selling something rare.