The death of SaaS and YoY category rising from its ashes

The founders celebrating the death of SaaS are about to miss the 108% YoY category rising from its ashes. ## Per-seat is the business model dying. Software demand is not. AI-native SaaS spending grew 108% year-over-year in 2026 while traditional SaaS grew 8%, according to the [BetterCloud/Zylo 2026 SaaS Management Index](https://www.bettercloud.com/monitor/saas-industry/). Agents need APIs, auth, data pipelines, and compute running around the clock. They don't need dashboards. They don't take PTO. IDC forecasts 1,000x growth in AI inference demand by 2027. The API call volume agents generate creates a software consumption floor with no human ceiling. The per-seat model collapses not because software becomes less valuable but because the unit of consumption is no longer a person. ## The incumbents proved the model doesn't die. It mutates. Salesforce and Intercom are the canonical case studies. Intercom pivoted Fin AI to $0.99 per resolution and hit $100M+ ARR. Salesforce ran three simultaneous pricing tiers: $2/conversation (outcome), $0.10/action via Flex Credits (consumption), and $125/user/month (per-seat). The company didn't kill per-seat. It layered on top of it. Intercom president Archana Agrawal: "Outcome-based pricing exposed every weak link in their organization — sales could no longer optimize for licenses, CS could no longer hide behind usage, RevOps had to forecast outcomes, and the product had to actually work consistently." The shift to outcome pricing isn't just a billing change. It's an organizational pressure test that most incumbents will fail slowly. ## The window is narrow. Gartner projects 40% of enterprise SaaS spend shifts to usage or outcome-based pricing by 2030. That sounds like runway. It isn't. Enterprise procurement cycles run 12-18 months. Infrastructure agents will subscribe to in 2028 needs to be built and under contract by late 2026. Incumbents are hedging: hybrid models, fixed base plus variable consumption. That gap is where agent-native infrastructure wins. The specs are clear: auth that speaks OAuth and API keys, not SSO and SAML; pricing that meters by action or outcome, not seat; APIs documented for machines, not humans. ## Velocity is not universality. Gartner projects per-seat revenue share declining from 21% to 15% by 2030. That means 85% of enterprise SaaS revenue is still not outcome-based in four years. Salesforce's $2/conversation pricing priced out SMBs. Only 3,000 of 5,000 initial Agentforce deals were paid. The $2T that evaporated from legacy SaaS valuations wasn't from software becoming worthless. It was from the market repricing the ceiling: from "everyone who uses this software" to "everyone who uses this software, knowing agents will handle volume that doesn't require human judgment." That's real compression. It's not apocalypse. The founders building agent wrappers on canceled SaaS subscriptions are building in the wrong direction. The opportunity is infrastructure: API-first products agents subscribe to, auth and billing primitives agents consume, the data layer feeding agent memory. That category grew 108% last year. IDC expects agents to outnumber human users of enterprise software before 2027. The death of SaaS is the story. The birth of agent-native SaaS is the trade.