"The Big Short's" Michael Burry Just Threw Cold Water on SaaS Armageddon. 3 Software Stocks He's Buying

The immediate read-through is not “AI is harmless,” but that the market has likely generalized one narrow margin-erosion narrative across businesses with very different moat structures. The cleaner setup is in scaled, transaction-linked software and data platforms where distribution, switching costs, and embedded workflows matter more than code-generation speed; that favors PYPL, CRM, and MSCI over smaller, feature-rich software names. A second-order effect is that if private-credit stress tied to software borrowers has been a hidden driver of de-risking, any stabilization in credit spreads can create a fast multiple rebound in these same names over the next 1–3 quarters.

PYPL looks like the most asymmetric because the market is still pricing it as a secular loser despite a balance-sheet-light, network-effect business with operating leverage if management simply avoids self-inflicted dilution. The key issue is not competition from AI, but wallet-share erosion from platform-level checkout integrations; that means the stock’s rerating depends more on stabilizing branded checkout volume and take rate discipline than on “winning AI.” If those metrics flatten rather than improve, the stock can stay cheap for longer, but downside should be limited if execution remains steady.

CRM is the higher-quality but less obvious catalyst-driven long: the risk is not disruption, it’s a slower growth slope that compresses the multiple until AI monetization becomes visible. Agentic automation can actually increase expansion revenue if it deepens admin dependence and raises switching costs, but that payoff is likely a 6–18 month story, not a next-quarter story. MSCI is the most underappreciated because the market tends to punish data/analytics names on generic AI fears, yet the real risk is more about pricing power and workflow bundling than product replication; if management can show AI is improving analyst productivity instead of replacing demand, the multiple gap versus its own history should close.