SaaS is Dead. What Comes Next?
A few weeks ago I noticed something strange while scrolling through my feed. Not one story, but four, all pointing at the same thing from different angles.
TinySeed, one of the most respected SaaS accelerators out there, quietly skipped opening applications for their next cohort. Not because they ran out of interested founders. Because they're genuinely unsure what qualifies as a defensible SaaS business anymore; if the people whose entire job is picking winning SaaS startups are pausing to rethink the criteria, that's not a minor blip.
Around the same time, two solo builders made more money doing the opposite of building software than most SaaS founders make building it. Marc Lou raised $112,000 auctioning off ad space on his own body for a HYROX race, literal skin as inventory. Jonathan Wilke built outbid.lol, a website that's basically a leaderboard you pay to sit on top of, in a few hours over a weekend, and it's pulled in over $260,000 since. No onboarding flow. No churn dashboard. No feature roadmap. Just an idea, shipped fast, monetized directly.
And then the big one. Bending Spoons, the Italian company known for buying up struggling tech brands and squeezing them into shape, picked up Airtable for $1.285 billion. Airtable was valued at $11 billion in 2021. A few weeks later, Bending Spoons did it again with Miro, paying $1.355 billion for a company that was worth $17.5 billion in 2022. These aren't failed startups. These are category-defining products with hundreds of millions in revenue, sold for a fraction of what they were once worth.
Put those together, and you get a pretty uncomfortable question: is the SaaS model itself starting to break?
Here's what's actually happening
I don't think SaaS is dead in the sense that software stops being valuable. What's dying is the assumption that made SaaS a safe bet: build a decent product, wrap it in a subscription, and defensibility takes care of itself over time.
That assumption doesn't hold anymore. AI has made building competent software fast and cheap. If your product is a UI on top of a database with some workflow logic, someone can clone the core of it in a weekend now, the way Wilke built outbid.lol in three hours. When the cost of building drops that low, "we built something useful" stops being a moat. It's the starting line.
TinySeed pausing applications is them admitting that publicly. If most of the pitches coming in don't have anything protecting them beyond "we shipped it first," that's not a fundable moat in 2026; it's a head start that AI erases in a few months.
The Bending Spoons deals tell the second half of the story. Airtable and Miro didn't lose their user bases. Airtable is still pulling around $480 million in annual recurring revenue. Miro has close to 4 million paying users. What collapsed was the multiple investors were willing to pay for that revenue, because the market stopped believing recurring revenue alone justifies a premium valuation when the underlying product isn't hard to replicate anymore.
And the indie hacker stuff, the body ads, the pay-to-rank leaderboards, that's the flip side. When trust in "subscribe and we'll keep improving this for you" erodes, people start reaching for models that are blunt, transparent, and immediate. You pay, you get the thing, done. No recurring relationship required.
So what comes next
If thin SaaS is getting commoditized, what survives is anything with a real moat that AI can't shortcut. That's usually one of a few things: proprietary data nobody else has access to, genuine network effects, deep integration into a workflow that's painful to rip out, or a brand and distribution advantage that took years to build. Feature lists don't make that cut anymore. Data, relationships, and distribution do.
I also think the pricing model itself is up for reconsideration. Subscriptions made sense when the ongoing engineering and support cost justified charging every month. But if a tool's core value is delivering a dataset or doing one job well, a one-time payment is often the more honest exchange, and increasingly what buyers actually want. Fewer people want another card on file for a tool they'll open twice a month.
What I'm doing about it
This is exactly the bet I'm making with Adamlead, the contact data platform I'm building. Instead of trying to be a smaller Apollo.io with more features, I picked one specific, defensible thing: verified contact and funding data on startups that just raised money. That data itself is the moat, not the interface sitting on top of it. Anyone can clone a dashboard. Building and maintaining a verified dataset of funded startups is a lot harder to copy over a weekend.
It's also why I've been testing one-time pricing on Adamlead rather than defaulting straight to a subscription-only model. If the value is "here's the data you need right now," charging once for access to it is a more honest transaction than asking someone to keep paying every month for something they might only need occasionally. I'd rather build something people are glad they paid for once than something they quietly forget to cancel.
I don't think this makes me immune to everything I just described. But I'd rather build on a real moat and a pricing model that matches how people actually want to buy in 2026 than build another thin SaaS wrapper and hope the subscription model keeps propping it up.
SaaS isn't dead. The version of SaaS that gets by on features and a monthly invoice is.