Opinion
I skipped the SaaS funeral, you should too
"In the short term, uncertainty is pushing multiples down. But in the longer term, as the “build-it-yourself” myth fades and pricing shifts from seat-based to value-based, valuations will move back in line with the real demand behind SaaS products. That’s why SaaS isn’t going anywhere," writes Yaniv Jacobi, Co-Founder and Managing Partner at Horizon Capital.
It’s become a bit of a sport lately to declare that "SaaS is dead". It’s a sharp, punchy headline that makes whoever says it sound like they’re already living in 2030. But the more people repeat it, the less accurate it gets. What we are witnessing isn't the death of SaaS; it’s a phase shift. The form is changing, the mechanics are evolving, and pricing models are being broken and rebuilt - but the core concept of Software as a Service isn’t going anywhere. On the contrary, we are entering an era were building a larger, faster software company is easier than it was a decade ago. And that is exactly why SaaS is about to be everywhere.
Let’s start from the top. When headlines scream "SaaS is dead," do they really mean the era of Software as a Service is over? Are they suggesting that organizations no longer consume software through a provider that maintains, secures, improves, and supports it? Not quite.
Neither AI breakthroughs - which allow anyone to build a "custom tool" - nor a regression to a world of local installations - manual upgrades, broken versions, and total dependence on internal IT teams are coming to save us. Organizational reality isn't heading in that direction. It’s moving the opposite way: more service, more provider accountability, more automation, more integration, and higher standards. This isn’t the "end of SaaS"; it’s its maturation.
Market Uncertainty is a Repricing, Not a Collapse in Demand
The primary reason the "SaaS is dead" narrative feels credible is that markets, especially public ones, have undergone painful correction. Valuations dropped, multiples shifted, and the narrative became "the category is over." But a drop in multiples isn’t proof that a category has lost its utility. It simply means the market isn’t sure how to price the future.
This is where AI enters like a thick fog, raising two heavy questions: How easy will it be to commoditize features? And can companies maintain their pricing power? Without clear answers, investors are hitting the brakes. But inside this fog, the dry facts remain: the SaaS market is massive and continues to grow, even if the pace has shifted.
The point isn't the exact growth percentage. The point is that underlying demand remains. Companies are reporting lower-than-expected churn rates, and the actual workload handled by software is increasing. Organizations still need CRMs, security tools, operational systems, data management, and automation. What’s changing is the distribution, the integration, and the business package—not the need itself.
The "Build-It-Yourself" Myth
The second argument goes like this: If we can generate anything today using models and agentic layers, why would companies buy? Why not just build? This leads to a better question: Can organizations customize more deeply than before? Yes. Do they want to build and maintain the core of every business system themselves? Almost always, no.
If I’m a pharmaceutical company, I have zero reason to turn into a software house just to build an internal CRM. I might want unique automations and integrations tailored to my business, but I don’t want to own the infrastructure, the continuous R&D, the maintenance, or the security layers around sensitive data. I want a SaaS product that allows for more flexibility.
Then there’s the stuff no one likes to put in a tweet: security, compliance, regulations, audits, and permissions. Any serious enterprise handling customer names, contracts, and financials won't "play" with their data. They won’t risk an unmonitored, unmanaged environment without logs or role-based access control. They would prefer a product that comes with built-in protection, protocols, and a roadmap.
And even if we set security aside: What happens when there’s a bug? Who’s responsible for the fix? Who provides support? One of the greatest promises of SaaS is that you aren't buying "code"; you’re buying a living system that someone is committed to operating and improving for you. AI doesn’t eliminate this need; it only raises the expectation for the system to improve faster.
From "Seat-Based" to "Value-Based" Pricing
The third argument is where the real shift is happening: pricing. Seat-based SaaS worked for years because it was simple: more employees meant more licenses, and growth was easy to model. But in the world of automation and AI agents, output won’t scale linearly with headcount. Teams will do the same work with fewer people, and if revenue is tied to seats, that math fuels the simplistic logic: "Fewer users = lower revenues = SaaS is dead".
But that conclusion undervalues what the product actually delivers and ignores its continued relevance. Instead, we are witnessing a clear move toward usage-based pricing, volume-based models, or "value-based" (per outcome) pricing. It won't matter if two employees use the system or fifty; if the system saved hours, mitigated risk, or increased conversion, it should be priced accordingly.
So yes, in the short term, uncertainty is pushing multiples down. But in the longer term, as the “build-it-yourself” myth fades and pricing shifts from seat-based to value-based, valuations will move back in line with the real demand behind SaaS products. That’s why SaaS isn’t going anywhere. The category is simply maturing, and we are about to see a new generation of SaaS companies built for this reality.
Yaniv Jacobi, Co-Founder and Managing Partner at Horizon Capital.
