← Back to writing

The SaaS Pricing Crisis

SaaSAI

Earlier this month, approximately $285 billion in SaaS market value evaporated. Per-seat licensing—the model that built the modern software industry—is dying. And this is just the beginning.

The reason is straightforward: AI agents are replacing human seats. When one agent can do the work of five people, charging per seat doesn't make sense anymore. The math breaks for buyers, and the revenue breaks for sellers.

But the real story isn't about pricing models. It's about a fundamental shift in who—or what—uses software.

Software Is Going to Look Completely Different

In two years, the way we think about software will be unrecognizable. We're building for the AI economy now, whether we realize it or not.

Here's what I mean: machines need to be able to access, evaluate, and purchase software. Agents are going to spin up exactly what software you—or they—need, on demand. The transaction layer between AI agents and software services is going to be one of the most critical pieces of infrastructure built in the next few years.

"We're not just changing how software is priced. We're changing who buys it."

Think about it: an AI agent managing your business operations doesn't browse a SaaS marketplace and compare feature lists. It needs an API it can call, a trust layer it can verify, and a pricing model that maps to consumption—not headcount.

The Trust and Transaction Layer

This is the part most people aren't talking about yet. When agents become the primary consumers of software, you need infrastructure for:

  • Machine-readable trust. How does an agent know this service is reliable? You need reputation systems, uptime guarantees, and verification that agents can evaluate programmatically.
  • Programmatic transactions. Agents need to negotiate, purchase, and consume services without human intervention. The checkout flow is an API call, not a web form.
  • Usage-based everything. When software is consumed by agents, pricing has to be granular. Per-request, per-outcome, per-credit. The unit of value shifts from "a seat" to "a result."

What This Means for Builders

Whether this is a threat or an opportunity depends entirely on what you're building.

If you're building a traditional SaaS product that relies on per-seat revenue from human users, you need to start thinking about your transition now. Not because the market is panicking—but because your customers' agents are going to start asking why they're paying for ten seats when one agent handles the work.

If you're building new products, you have an advantage: you can architect for the AI economy from day one. That means:

  1. API-first, always. Your product's primary interface might be an API consumed by agents, not a dashboard used by humans.
  2. Price for consumption. Tie pricing to value delivered, not access granted. Credits, usage, outcomes.
  3. Build for machine discovery. Make your product easy for agents to find, evaluate, and integrate with. This is the new SEO.
  4. Invest in trust infrastructure. Reliability, transparency, and verifiable performance become competitive advantages when your buyer is an algorithm.

This Is Just the Beginning

The $285 billion correction wasn't an overreaction. If anything, the market hasn't fully priced in what's coming. We're in the early innings of a shift from human-centric software to agent-centric software, and the business models, interfaces, and infrastructure haven't caught up yet.

The builders who thrive will be the ones who stop optimizing for how software works today and start building for how it will work when agents are the primary users. That future is closer than most people think.