The short
- Shift: Enterprise SaaS prices rose 11–18% YoY—highest since 2019.
- Cause: Cloud vendors pass AI compute and storage inflation downstream.
- Response: CFOs trim “tool sprawl” and audit underused subscriptions.
- Countertrend: Open-source and self-hosted tools surge 26% in adoption.
- Emotion: The cloud isn’t weightless anymore—it bills by the breath.
The decade that made software expensive again
It began as freedom: one-click deployment, per-seat pricing, and auto-updates that never broke your weekend. But over time, those $8-per-user apps multiplied. The average midsize company now pays for 107 SaaS tools. Only 56 are used weekly.
The pandemic turbocharged this clutter—remote work needed collaboration suites, ticketing tools, HR dashboards, and analytics everywhere. But as rates rose and AI workloads exploded, the same vendors quietly rewrote price sheets.
The new price landscape
| Vendor | 3-Year Price Change (%) | Churn (%) | Key Driver |
|---|---|---|---|
| Microsoft 365 | +15 | 3.4 | AI Copilot bundles |
| Atlassian | +21 | 4.1 | Cloud migration costs |
| Salesforce | +18 | 3.9 | Data storage expansion |
| Zoom | +9 | 5.5 | Flat usage, margin defense |
| Adobe | +14 | 2.6 | GenAI integration costs |
Data: Q3–Q4 2025 renewal analysis, enterprise billing records (North America, EMEA, India).
“Do we still need this?”—the CFO awakening
“There’s an app for that” became “there’s an invoice for that.” Companies are running line-by-line audits. Some now maintain internal “stack scorecards” — mapping overlap and uptime against ROI.
Enterprises are cutting back on niche analytics and secondary collaboration apps, often consolidating 10 tools into 3. Procurement teams are learning to negotiate like hedge funds — pairing usage data with per-seat renegotiations.
How AI broke the old economics
GenAI sounded like a profit windfall — but inference workloads are compute-hungry. Training a chatbot for internal support might cost $250K in cloud credits annually. Vendors are passing that load to customers as “AI convenience fees.”
AI-first pricing is now bundled into everything — even if users don’t enable it. The irony: companies are paying for features they turn off to save money.
Who wins, who exits
- Winners: Open-source tools (e.g., Metabase, Plausible) and regional SaaS players undercutting hyperscalers.
- Losers: Mid-tier collaboration apps with unclear ROI and overlapping functions.
- New entrants: “SaaS auditors” — startups offering savings-as-a-service, paid by recovered costs.
The future stack
In 2026, the pendulum swings toward simplicity: smaller stacks, higher stickiness, lower noise. The great correction isn’t anti-cloud — it’s anti-excess. After a decade of adding tools, the next edge is subtraction.
Software used to free time; now, it’s being asked to earn it back.