Here's the thing nobody on your board wants to say out loud: the business model that turned software into the most profitable industry on Earth is breaking.

Wall Street has a name for it now — the SaaSpocalypse. Between January and mid-February 2026, approximately $2 trillion in market capitalisation evaporated from the software sector. The iShares Expanded Tech-Software ETF (IGV) dropped 22% year-to-date, marking the steepest software selloff since the 2022 rate-hike cycle. Atlassian fell 35% after reporting its first-ever decline in enterprise seat count. Salesforce shed 28% despite still growing revenue. SAP plunged 16% on a disappointing cloud forecast. This isn't a rotation. It's a repricing of an entire industry's economics.

The trigger was straightforward: agentic AI crossed a threshold. When OpenAI launched Frontier — a full-stack orchestration layer for deploying autonomous agents that can navigate existing software interfaces — and Anthropic shipped Claude Cowork, the market suddenly understood that if one AI agent can handle the workload of 10 to 15 mid-level employees, organisations no longer need 10 to 15 Salesforce seats. The equation that built modern SaaS — more employees equals more seats equals more revenue — is running in reverse.

The data confirms this isn't hysteria. Atlassian's enterprise seat count declined for the first time in the company's history. Project management, task tracking, data entry, and customer logging — the bread-and-butter workflows of tools like Jira and Salesforce — are precisely the tasks AI agents automate most effectively. Gartner now predicts 35% of point-product SaaS tools will be replaced by AI agents by 2030. And 65% of SaaS vendors have already started bolting usage-based pricing onto their seat-based models, a tacit admission that the old meter is broken.

But the disruption cuts deeper than pricing. AI-native SaaS companies are running with gross margins of 50–65%, compared to the 78–85% that traditional SaaS enjoys. Every model invocation costs real money. Variable COGS per user runs at 20–40% of revenue versus less than 5% for traditional software. Bessemer Venture Partners found that scaling AI companies average just 25% gross margin. That's not a SaaS business — that's a services business wearing a software costume. Early-stage AI-native companies are targeting 25–40% gross margins and calling it acceptable. Even at maturity, 60% is the aspirational ceiling. The 80%+ margins that justified SaaS's premium multiples may never return for companies where inference is the product.

The retention numbers are equally sobering. Research from Growth Unhinged reveals that AI-native products selling for under $50 per month see just 23% gross revenue retention and 32% net revenue retention. Products in the $50–$249 range fare better at 45% GRR and 61% NRR, but that's still 15 percentage points worse than traditional B2B SaaS benchmarks. Only AI-native products priced above $250 per month — the ones solving genuinely complex enterprise problems — achieve retention metrics comparable to the SaaS companies we've benchmarked against for the past decade.

Founders navigating this have three strategic levers. First, smart model routing: sending 80% of requests to cheaper, smaller models and reserving expensive reasoning models for the 20% of cases that actually need them can reclaim 15–20 points of margin. Second, hybrid pricing: 92% of AI software companies now combine subscriptions with usage fees, because pure per-seat pricing leaves money on the table when agents are doing the work. Third, outcome-based billing: Zendesk's approach of charging per resolved ticket for AI agents while maintaining per-seat pricing for humans is an early template. The companies that figure out how to bill for value created rather than humans employed will win the next era.

The counter-argument deserves airtime. Not all SaaS is doomed — the global market is still forecast to hit $430 billion in 2026. Companies with deep data moats, network effects, and regulatory compliance advantages will survive and likely strengthen. The selective unbundling thesis suggests that commoditised point solutions die while differentiated platforms endure. And the current stock selloff may be overshooting, as it did in 2022. Wall Street has a habit of pricing in decade-long disruption in a single quarter.

But here's the founder takeaway: whether the SaaSpocalypse is a market overreaction or the beginning of a structural shift, the pricing and margin assumptions you built your company on are being stress-tested in real time. If your gross margin depends on per-seat pricing and your workflows are automatable, you're on the wrong side of this trade. The founders who survive will be the ones who treat pricing as a core product decision, who build margin discipline into their architecture through model routing and fine-tuning, and who find ways to capture the value AI creates rather than watching it compress their seat count. The era of software economics on autopilot is over.