Private equity's smartest money admits they can't price software companies anymore, and AI is why.
The Summary
- Apollo's PE co-head says AI is making software valuation fundamentally harder, not easier
- Traditional metrics for pricing software assets are breaking down as AI rewrites what software companies actually do
- Meanwhile, PE firms are still dealmaking through geopolitical chaos because "fortunes are made in volatility"
The Signal
David Sambur runs half of Apollo's private equity machine. When he says valuing software firms is getting harder because of AI, that's not a complaint. It's an admission that the old playbook is burning.
For decades, PE firms valued software companies using clean math: recurring revenue multiples, customer acquisition costs, churn rates. Predictable inputs, predictable outputs. But AI agents are scrambling the equation. Is a SaaS company worth more or less when its customers can deploy agents that do the same work for a fraction of the cost? What's the multiple on revenue that might evaporate when enterprises realize they're paying for human-speed software in an agent-speed world?
Apollo manages over $700 billion. They don't flinch at macroeconomic turbulence or war risk. Sambur's point about making fortunes in volatility is PE 101. But structural uncertainty about what a software company even is in three years? That's new terrain. The firms that figure out how to value software in the agent era will own the next decade of returns. The ones still using 2023 models will get cleaned out.
The Implication
If you're building or investing in software, stop optimizing for metrics that won't matter in 18 months. The question isn't "what's our ARR growth" anymore. It's "what part of our product can't be replaced by an agent, and how do we price that?" If Apollo is confused, everyone else is just pretending they're not.
Source: Bloomberg Tech