SaaS Headwinds

September 14, 2025

Is SaaS dead? No. But there are real headwinds, and all of them trace back to AI.

SaaS has been the dominant software delivery model for two decades. Salesforce, Workday, ServiceNow, Atlassian. The playbook is well understood: build a horizontal or vertical product, sell subscriptions, expand seats, layer on platform fees. It worked because building software was expensive and buying was almost always cheaper than building. That equation is changing, and the implications for future SaaS revenue are significant.

Niche Local Players Will Multiply

Consider a textile factory in Turkey that needs an ERP system. Five years ago, building a Turkey-specific textile ERP was not a viable business. The market was too small to justify the development cost. So the factory bought SAP or a generic cloud ERP, paid for localization consultants, and lived with a system that was 60 percent relevant to how they actually operated.

AI coding tools have changed the math. The cost of building that niche ERP has dropped by an order of magnitude. A small team, or even a single technical founder, can now build a textile-specific system that handles Turkish tax compliance, local supply chain workflows, and industry-specific production scheduling. The market size has not changed, but the minimum viable investment to serve it has collapsed.

This pattern will repeat across thousands of niches. Regional, industry-specific, and workflow-specific software that was never worth building at the old cost structure is now viable. Many of these businesses will be bootstrapped. They do not need venture capital because the build cost is low enough to fund from revenue. Each one individually is small. Collectively, they eat significant share from horizontal SaaS incumbents who served these niches poorly but without competition.

Build vs. Buy Is Shifting Toward Build

The core SaaS value proposition is that buying is cheaper than building. For most enterprise software needs, this has been true for 20 years. It is becoming less true every quarter.

AI-assisted development is compressing build timelines and reducing the engineering headcount required to ship production software. A system that would have required a team of eight engineers and six months can now be built by three engineers in two months. The gap between "buy a SaaS product" and "build exactly what we need" is narrowing.

This does not need to be a wholesale shift to matter. If the build-vs-buy decision moves from 90/10 in favor of buying to 85/15, that is a five-percentage-point reduction in the addressable market for SaaS companies. For a category doing $500 billion in annual revenue, five points is $25 billion that shifts from SaaS subscriptions to internal engineering budgets. The shift does not need to be dramatic to be material.

Internal Teams Will Build More

This is related to the build-vs-buy shift but distinct. Even companies that continue buying SaaS products will build more internal tooling around and between those products.

Today, a typical enterprise has dozens of SaaS subscriptions and a web of integrations connecting them. Many of these integrations are themselves SaaS products: Zapier, Workato, MuleSoft. When internal teams can build custom integrations, internal dashboards, and workflow automation faster and cheaper with AI tooling, the integration-layer SaaS products lose share first. But the primary SaaS products lose share too, because internal teams start replacing lightly-used SaaS tools with purpose-built alternatives that fit their exact workflow.

The employee who previously submitted a procurement request for a $30,000/year SaaS tool now builds a working version over a weekend. It is not as polished. It does not need to be. It does exactly what the team needs and nothing else.

The Combined Effect

Each of these headwinds is individually modest. Local niche players will not kill Salesforce. The build-vs-buy shift will not empty out AWS Marketplace overnight. Internal teams will not replace every SaaS subscription with a homegrown tool. But all three forces are directionally aligned, all three are accelerating, and all three are driven by the same underlying cause: AI is making software cheaper to build.

For investors, the implication is straightforward. SaaS companies that are priced for 20-30 percent annual growth in perpetuity are carrying risk that the market has not fully priced in. The headwinds will not show up as a cliff. They will show up as a gradual compression of growth rates, higher churn in the mid-market, and slower seat expansion as internal alternatives absorb use cases that would have gone to SaaS vendors. The companies most exposed are horizontal SaaS products serving commodity workflows in markets where local or vertical alternatives can now be economically built.

SaaS is not dead. But the tailwind that carried it for two decades is weakening, and the new headwinds are just getting started.