Between February 4 and February 6, 2026, the SaaS industry lost $285 billion in market cap. Two days. The financial press called it the SaaSpocalypse.
The headlines focused on the market cap.
But they didn’t write about the cause…
Wall Street has already voted
Over the past 12 months, the four canonical seat-priced enterprise software stocks (Salesforce, Adobe, ServiceNow, and Atlassian) shed roughly a third to two-thirds of their market cap.
Atlassian alone is down 69%.
The S&P 500 was roughly flat over the same window.
The cause?
Seat compression.
AI agents need fewer humans, fewer humans need fewer seats, fewer seats means a structurally smaller pie for every per-seat business.
- Atlassian reported its first-ever decline in enterprise seat counts and cut 10% of its global workforce to "self-fund AI investments."
- Workday cut 8.5%.
- Oracle cut between 10,000 and 30,000 jobs.
Q1 2026 alone saw roughly 80,000 tech layoffs, about half of them explicitly AI-attributed.
The slowest growth in a decade
Even before AI, the model was running out of road. Median public-SaaS revenue growth has fallen every year since the 2021 peak: 33% → 24% → 17% → 16% → 12.2%.
Sixteen straight quarters of deceleration.
The cause of death
Across customer support (the most quantifiable AI workflow on Earth) modern AI agents resolve 55-70% of tickets without a human. Agentic deployments routinely hit 70-85%. Top performers clear 86%.
HubSpot's Breeze Customer Agent reports 65% resolution across 8,000 customers. Intercom's Fin charges per resolution because resolutions are now the unit of value.
For 25 years, every business software company has charged you for a seat. You paid whether the work got done or not.
The seat made sense when humans were doing the work. AI agents now do most of it.
The receipts
A thought experiment using public Zendesk pricing:
A team running 1,000 support tickets a month, paying the public Zendesk Suite price ($115-155/agent/month), the AI Copilot add-on ($50/agent/month), and AI resolution fees ($1.50-2 each), is at $2,000-3,000 a month before a single ticket gets resolved.
The same workload at $0.75 per AI resolution with no platform fee costs roughly $462.
The incumbents see this math as clearly as anyone.
- Salesforce just went "headless," making MCP available without a Salesforce seat.
- ServiceNow says 50% of new business revenue is now non-seat consumption.
- Anthropic restructured Claude Enterprise from $40-200/seat flat to $20/seat plus usage.
- HubSpot moved Customer Agent to $0.50 per resolved conversation.
- SAP announced a wholesale shift to AI consumption pricing.
The 25-year pattern
Every era of business software has ended with a pricing rebellion vs a feature war.
Oracle and Siebel weren't out-engineered in the early 2000s. They were out-priced. Marc Benioff stood outside the Siebel user conference with a "No Software" sign and killed the perpetual license model with the seat subscription.
Twenty-five years later, his own company is on the wrong side of the same protest.
Tomasz Tunguz called it in July 2024:
"Salesforce made famous the No-Software mantra competing on pricing. Perhaps usage or performance pricing will be the catalyst for a new era of upstarts displacing incumbents. Maybe we'll see a No-SaaS rebel replicate Marc Benioff's playbook."
Kyle Poyar, this April:
"You can't just sell seats anymore."
Jason Lemkin, February:
"When AI can do the work of multiple humans, you need fewer humans. When you need fewer humans, you need fewer seats. This is the quiet killer."
So what comes next?
Wall Street is repricing every per-seat company on Earth.
The smart money is converging on outcome-based pricing as the answer. The incumbents are quietly restructuring their pricing pages and hoping nobody notices.
You can't bolt a new pricing model onto a 25-year-old foundation. The legacy SaaS giants dipping a toe into outcome-based pricing are about to find that out the hard way.
This Thursday, I'll lay out what actually comes next (and why every "outcome-based" bolt-on you've seen so far is missing the point).
Cheers,