Private equity's long romance with software is fracturing. The sector's SaaS Spocalypse has laid bare how vulnerable recurring-revenue models look when artificial intelligence agents can replicate code, automate workflows, and shrink the value of seat-based licensing. General partners who built portfolios around predictable ARR growth now confront a buyer universe questioning whether yesterday's moats hold water.

Limited partners are pressing firms to demonstrate they can distinguish between AI-native platforms and legacy vendors bolting on copilots. The diligence burden has shifted: buyers no longer accept retention metrics at face value when a competitor's agent can onboard a customer in hours rather than quarters. Valuation multiples that peaked in 2021 now require re-underwriting.

The Matrix line — "Never send a human to do a machine's job" — reads less like fiction and more like an investment thesis. Autonomous agents threaten to collapse the services layer that padded software margins, turning implementation into a commodity. Firms that treated AI as a feature rather than an architectural shift are watching their thesis evaporate in real time.

The next vintage of software deals will be written by managers who understand that defensibility now sits in proprietary data, workflow depth, and switching costs that agents cannot easily replicate. Everything else is rent.