The Death of Software Equity: Why Companies Are Rethinking the Build vs Buy Debate

Scroll through LinkedIn and you’ll find two flavors of the “subscriptions vs. owning” debate. Consumers venting about streaming price hikes and BMW charging extra for heated seats. And SaaS founders celebrating their Annual Recurring Revenue multiples like it’s a personality trait.

Nobody’s talking about the number that should actually keep a CFO up at night.

The Operating Expense illusion

Ten years ago, the pitch made sense. Stop sinking capital into servers and licenses you own outright. Rent the software, stay agile, let OPEX absorb the cost instead of CapEx. For a while, that trade genuinely paid off.

It doesn’t anymore. It’s turned into a quiet, compounding tax.

Enterprise SaaS spending now runs $1,370 per employee annually, up 55% since 2021. And here’s the part that should really sting: as much as a quarter of those licenses just sit there, completely unused.

You’re not paying for tools. You’re paying rent on shelfware.

The misaligned multiples

There’s a structural reason vendors push subscriptions so hard, and it has nothing to do with what’s best for you.

Public SaaS companies trade at 2.0x to 3.5x Enterprise Value-to-Revenue, well above what perpetual license software companies command. That gap isn’t an accident. It’s the entire business model. Recurring revenue makes a vendor’s cap table look better, so vendors are financially incentivized to keep you renting forever, whether or not it’s the right call for your balance sheet.

You’re not the customer in that equation. You’re the annuity.

Companies Are Ditching the Cloud

This is where it gets interesting, because companies are starting to walk away.

Andreessen Horowitz’s now-famous “Trillion Dollar Paradox” research found that cloud and SaaS infrastructure can cut a tech company’s gross margins in half. In response, 72% of enterprises spending over $2 million a year on cloud are repatriating workloads, and they’re seeing 45% to 60% cost reductions when they do.

That’s not a rounding error. That’s half your infrastructure spend, recovered, just by owning instead of renting.

And it’s not only a cost story. Harvard Business School research led by Elie Ofek found that nearly 75% of companies rushed into subscription models chasing predictable revenue, but over 70% of buyers now report subscription fatigue. Churn is climbing. Retention is decaying. The model that was supposed to be a sure thing isn’t as sure as it used to be.

Add in an AI era where compute keeps getting cheaper and code generation keeps getting faster, and self-hosting stops being a heroic engineering lift. It becomes a realistic option again.

The own-your-core framework

None of this means “cancel every subscription and build everything in-house.” That’s its own kind of expensive mistake. What it means is drawing an honest line.

Rent for utility: Tools that evolve fast, where you benefit from someone else’s constant iteration. Third-party services where the value is in staying current, not in owning the code.

Buy and own for sovereignty: The stuff that’s actually core to how your business runs. Mission-critical databases, internal workflows, data-heavy engines you’ll depend on for a decade. These aren’t utilities. They’re assets, or they should be.

Every dollar you spend on the second category through a subscription is a dollar that builds someone else’s balance sheet instead of yours.

Your move

Here’s the exercise worth doing this week: take your top three SaaS tools and calculate your five-year cumulative spend. Now compare that to what a perpetual license or self-hosted deployment would’ve cost.

At the end of five years of renting, what do you actually own?

If the answer is “nothing,” you don’t have a software strategy. You have a landlord.