Software stocks trading at a discount to the S&P 500 is a first in modern market history. This didn't happen during the dot-com crash, the 2008 financial crisis, or the 2022 rate hikes.
What Is This?
@darraghog nailed it on X: "The 2026 SaaS crash is real. But it's not because AI agents are going to replace Salesforce next quarter. It's because AI is eating the budget." That's exactly what's happening.
The numbers are staggering. According to SaaStr's Jason Lemkin, software forward P/E multiples collapsed from 84.1x in 2020 to 22.7x in March 2026. The IGV (iShares Software ETF) is down 21%+ YTD, and roughly 30% from its September 2025 peak. That's $2 trillion in market cap gone.
Forbes calls it "SaaSpocalypse." The trading pattern is "get me out" style selling. Hedge funds shorting software stocks have made $24 billion.
Two forces are at work. First, AI is replacing seats. When one AI agent does the work of five people, enterprises buy 100 seats instead of 500. Salesforce growth fell from 25% to 8%, HubSpot from 47% to 17%, Snowflake from 106% to 24%.
Second, AI itself is consuming the budget. According to SaaS management firm Zylo, AI application spending jumped 400% YoY at companies with 10,000+ employees. Existing SaaS budgets are flowing to Anthropic and OpenAI — SaaStr estimates up to 70% of the software slowdown may be driven by this budget shift.
What's Actually Different?
Software commanded a market premium for two decades for good reason: 70-80%+ gross margins, predictable recurring revenue, negative net revenue churn. A dollar of new software revenue costs almost nothing to deliver.
This crisis is structurally different from previous ones.
| Previous Crises | 2026 SaaS Crisis | |
|---|---|---|
| Cause | Speculation / macro shocks | Business model under question |
| Nature | Temporary valuation reset | Structural revenue model shift |
| Recovery condition | Economy recovers → stocks recover | Seat-based model must survive |
| Software vs S&P | Always traded at premium | First-ever discount to market |
Thoma Bravo's Orlando Bravo — who's spent 20+ years buying and building software businesses — publicly said this month that valuation decreases for some AI-disrupted software companies are "very warranted." That's a significant statement from someone at his level.
Pricing models are shifting fast. Per-seat pricing dropped from 21% to 15% of SaaS companies in 12 months, while hybrid pricing surged from 27% to 41%. The transition to usage-based models like Salesforce's Flex Credits and Microsoft's Copilot Credits is accelerating.
How to Navigate This
- Audit your SaaS portfolio
Average enterprise SaaS apps dropped from 112 to 106. 82% of companies are reducing vendors. Cleaning up unused seats and duplicate tools frees budget for AI tools. - Renegotiate seat-based contracts
Renewals are the only cost reduction window. Rightsize licenses, negotiate hybrid or usage-based models. SaaS vendors are desperate to retain customers right now. - Track shadow AI spending
Employees expensing ChatGPT and Claude subscriptions individually is surging. Without centralized management, budget volatility and security risks increase. - If you run a SaaS: deeply integrate AI
Rebranding as "AI-powered" doesn't work. Use AI agents to reduce cost of complex business processes and show quantifiable ROI. BCG found CEOs spending 6+ hours/week upskilling on AI are 12x more likely to be in the top 5%.
Bull vs Bear: Two Scenarios
Buying opportunity: At 22x forward P/E, companies with data moats, enterprise relationships, and 10-20 years of customer lock-in look cheap if growth stabilizes.
Structural risk: If seat expansion as a growth engine disappears and pricing power weakens, even 22x may be too high for decelerating businesses.
