SaaS Is Dead. Here's What Businesses Are Building Instead.

There's a quiet realization spreading through businesses right now, and it usually starts with an invoice.

Maybe it's the CRM renewal that came in 30 percent higher than last year with an AI assistant nobody asked for bundled into the price. Maybe it's the project management tool with 40 seats provisioned and 11 people actually logging in. Maybe it's the moment a founder opens their expense report, counts fourteen separate software subscriptions, and realizes they couldn't explain what half of them do.

The realization is this: the way businesses buy software stopped making sense, and almost nobody noticed because it happened one auto-renewal at a time.

For twenty years, SaaS was the obvious answer. You had a problem, someone had already built the solution, and you paid a monthly fee to use it. Building your own software was absurd. It cost six figures, took a year, and required hiring engineers you couldn't afford and didn't know how to manage. Renting was rational. Everyone rented.

That logic is dead. Not dying, not evolving, not entering a new chapter. Dead. The economics that made SaaS the default choice for every business problem have inverted, and the businesses that see it first are going to spend the next decade running leaner, moving faster, and keeping money that their competitors are still mailing to Salesforce.

The math stopped working

Start with what businesses are actually getting for their SaaS spend, because the numbers are genuinely hard to believe.

According to Zylo's 2026 SaaS Management Index, the average company now runs roughly 275 SaaS applications and wastes about $21 million a year on licenses nobody uses. Fifty-three percent of SaaS licenses sit completely idle. Vertice's Q1 2026 data, drawn from over $30 billion in processed software spend, found that 66 percent of all SaaS licenses are either untouched or surplus to actual requirements. Fifteen percent of applications show zero activity at all. They were purchased, integrated, and forgotten, kept alive only by auto-renewal clauses and a lack of anyone paying attention.

Read that again. Two thirds of the software the average business pays for delivers nothing. This isn't a rounding error or a procurement hiccup. It's the defining characteristic of how the industry works.

And the prices are going up anyway. PricePulse tracked more than 90 SaaS tools through the first half of 2026 and found the average company saw a 34 percent cost increase, at a time when general inflation ran around 3 percent. SaaS pricing is inflating at ten times the rate of everything else, driven by AI features bundled into plans whether you want them or not, seat count audits, and the quiet elimination of legacy pricing tiers. Seventy-nine percent of IT leaders encountered price increases at renewal in the past twelve months.

So the average business is paying more every year for software where most of the licenses go unused, most of the features go untouched, and none of the workflow actually matches how they operate. They adapt their business to the tool, because the alternative was always too expensive.

Was. That word is doing a lot of work, and it's the whole point of this post.

The Klarna moment

In late 2024, Klarna's CEO Sebastian Siemiatkowski announced on an investor call that the company was shutting down Salesforce and Workday, two of the biggest names in enterprise software, and replacing them with internally built systems powered by AI. The reaction from the industry was mostly disbelief. Marc Benioff himself went on a podcast demanding to know how Klarna could possibly manage its customer and employee data without his product.

Plenty of commentators rushed to point out the caveats, and some of them were fair. Klarna is a large fintech with real engineering talent. Part of what they did was consolidation and standardization, not pure AI magic. This wasn't something a 40-person logistics company could copy on Monday morning.

But the caveats missed what actually mattered. The significance of Klarna wasn't the specifics of their tech stack. It was that a serious company looked at the two most entrenched pieces of enterprise software on the planet, ran the math on building versus renting, and building won. Five years earlier, that math was not close. The fact that it was even a conversation meant something fundamental had shifted underneath the entire industry.

What shifted is that the cost of creating software collapsed.

What actually changed

GitHub's most recent State of the Developer research reports that 92 percent of developers now use AI coding tools, and roughly 41 percent of all code written last year was AI generated, a share projected to pass half in high-adoption teams by the end of 2026. McKinsey's research puts productivity gains for teams that build with AI in their development process at 35 to 45 percent. These aren't marketing numbers from tool vendors. This is the documented, measured reality of how software gets made now.

What used to take a six-person team six months can now be done by two people in six weeks. What used to require a $250,000 budget can be scoped in phases starting at a fraction of that. The work didn't get easier in the sense that judgment stopped mattering. It got faster in the sense that the expensive, repetitive, boilerplate-heavy middle of every software project largely evaporated.

And businesses have noticed. Retool surveyed 817 builders for its 2026 build-versus-buy report and found that 35 percent of teams have already replaced at least one purchased SaaS tool with something custom built, and 78 percent plan to build more this year. Every single SaaS category is under replacement pressure, with workflow automation and internal admin tools leading the way. Those are exactly the categories where the gap between what a generic tool offers and what a specific business needs is widest.

Meanwhile, the market itself delivered a verdict. Over two days in February 2026, the SaaS sector lost roughly $285 billion in market value while the broader S&P stayed nearly flat. Investors weren't panicking randomly. They were repricing a specific category of company: software vendors whose revenue depends on per-seat billing in a world where AI keeps reducing the number of seats businesses need. A Cruxy survey of 300 SaaS CEOs in April found that 97 percent plan to retire seat-based pricing within two years. When nearly every CEO in an industry says the core business model has an expiration date, believe them.

The per-seat tax was always the tell

It's worth sitting with why per-seat pricing is collapsing, because it reveals what SaaS actually was.

Per-seat pricing never had anything to do with the cost of delivering software. An additional user costs a SaaS vendor approximately nothing. The seat was simply the most convenient way to charge you more as you grew. It was a tax on your headcount, dressed up as a pricing model. When your business did well and hired people, your software bill went up, even though the software itself didn't change at all.

Businesses tolerated it because there was no alternative. And now watch what's happening as vendors scramble to replace it: AI credits, consumption tiers, per-resolution fees, hybrid models with a fixed base plus variable charges that make forecasting your own software budget an act of guesswork. Zylo found that 78 percent of IT leaders got hit with unexpected AI or consumption charges in the past year, and 61 percent cut planned projects because of price increases. The industry isn't fixing the misalignment between what you pay and what you get. It's reshuffling it.

There's a version of this that works out fine for you as a customer, where usage-based pricing genuinely aligns cost with value. But there's a simpler resolution available now that wasn't available before, which is to stop renting the software entirely.

What replaces it: custom micro software

Here's the future I'd put money on, because at 918 Studio we're already living inside it every week.

Businesses stop buying sprawling platforms with 400 features and start owning small, sharp, custom tools that do exactly what their operation needs and nothing else. Call it micro software. A quoting tool built around how your sales team actually quotes. A scheduling system that knows your crews, your territories, and your rules. A client portal that matches your process instead of forcing your process to match someone's product roadmap in San Francisco.

The objection used to be obvious: custom software is expensive, slow, and risky. That objection was true, and it isn't anymore. AI-assisted development changed the cost structure of building so dramatically that a focused custom tool now frequently costs less than a year or two of the subscriptions it replaces. And unlike the subscription, you own it. No renewal negotiations. No surprise AI surcharge. No per-seat tax on your growth. No feature bloat you're subsidizing for someone else's enterprise customers. The software is an asset on your side of the ledger instead of a rent payment on the other side.

There's a deeper advantage that shows up after the cost savings, and it's the one our clients talk about most. When your software matches your workflow instead of approximating it, the friction disappears. Nobody's exporting data from one tool to reshape it in a spreadsheet before pasting it into another tool. Nobody's maintaining a wiki page explaining the workarounds. Research on tool sprawl found that nearly one in five workers switches between apps more than 100 times per day. That's not a software cost. That's an attention cost, and it's larger than the invoices.

The honest caveats, because this post should survive contact with a skeptic

If you've read this far and you're skeptical, your skepticism probably takes one of two forms, and both deserve a straight answer.

The first is that most businesses can't build software, AI or not. This is correct. The tools that let anyone generate an app in an afternoon are real, and they're genuinely useful for prototypes and personal tools. They are not, by themselves, how you should run your payroll or store your customer data. Software that a business depends on needs architecture decisions, security, data governance, testing, and someone accountable when something breaks at 4 p.m. on a Friday. Sam Altman called the flood of disposable AI-generated apps the fast fashion era of software, and even Retool, whose entire business benefits from the build trend, warns that hundreds of ungoverned one-off tools scattered across departments just recreates SaaS sprawl with worse security. The skeptics are right that vibe coding is not a business strategy.

The second is that some software genuinely should stay bought. Also correct. Accounting, payments infrastructure, email: categories where compliance burden is enormous, the products are mature, and your needs really are generic. Nobody should custom build their general ledger. The build-versus-buy question was never all or nothing, and anyone telling you to rip out everything is selling something.

But notice what both caveats actually imply. They don't rescue the old model where businesses default to renting bloated platforms for everything. They point to a specific gap: businesses need custom software built and maintained to a professional standard, without the cost of hiring an internal engineering team. That gap is where the next decade of business software actually gets decided.

The small studio as your tech team

Which brings me to the model I think wins, and full disclosure, it's the model we bet our company on.

The future for most businesses isn't building software themselves, and it isn't renting it from vendors who bill by the seat. It's partnering with a small, AI-native development studio that operates the way an internal tech team would, at a fraction of the cost of actually hiring one.

The math here is straightforward. A single mid-level software engineer costs $130,000 to $180,000 a year fully loaded, and one engineer alone can't cover design, architecture, and operations anyway. A real internal team starts around half a million a year. That was always the barrier, and it's why businesses settled for renting. But a small studio working with AI-assisted development serves that same role across multiple clients, which means each client gets senior product thinking, design, and engineering for less than the price of one junior hire. Often for less than their current SaaS bill.

The word that matters in that description is partner, not vendor. A vendor ships a project and disappears. A partner knows your business, maintains what they build, extends it as you grow, and picks up the phone. It's the relationship businesses have always had with their accountant or their attorney, applied to software. You don't employ your CPA full time, but you'd never say you don't have one.

This is exactly how we've structured 918 Studio, and it's why we work in phases rather than monolithic projects. Phase one solves the sharpest problem and proves the value in weeks. Later phases extend from there, funded by results instead of promises. The client owns the software at every step. When AI collapsed the cost of building, it didn't just make software cheaper. It made this entire relationship model viable for businesses that could never afford it before, and that's the part of this shift that I think people still underestimate.

What this means if you run a business

You don't have to believe SaaS is dead as a category to act on any of this. The market will sort out which vendors survive. What you should believe, because the data says so plainly, is that the default has flipped.

The old default: a business problem appears, you search for software that solves it, you pick the least-bad option, you pay monthly forever, and you adapt your operation to the tool. The new default: you ask what a tool built exactly for this problem would look like, you find out what it costs to own it, and you're frequently surprised that owning costs less than renting.

So run the audit. Pull the list of every subscription you pay for, what it costs annually, how many seats you pay for versus how many get used, and how much of each tool your team actually touches. If your numbers look anything like the industry averages, half of what you're paying for is delivering nothing, and the rest is delivering a rough approximation of what you actually need at a price that rose 30 percent while you weren't looking.

Then ask a different question than the one you've been trained to ask. Not "what software should we buy?" but "what should our software be?" For twenty years only big companies got to ask that question. That era is over.

The businesses that figure this out first won't just save money. They'll run on software shaped precisely to how they work, owned outright, maintained by a partner who knows their operation, at a cost that would have sounded like a typo five years ago. Everyone else will keep paying the seat tax and calling it the cost of doing business.

It isn't anymore.

918 Studio is an AI-native product studio in Kansas City. We design and build custom software for founders and businesses in phases, and we stick around to run it like your internal tech team would. If you're staring at a SaaS renewal and wondering if there's a better way, there is. Talk to us.