The conversation covered token economics, data moats, and how AI is fundamentally reshaping technology due diligence for private equity.

A few key takeaways:

The end of zero marginal cost: Every AI workflow consumes tokens. Without tracking unit economics, what was once a high-margin SaaS business can quickly become a margin leak.

Mindless token burning: Throwing the largest, most expensive LLM at every task is inefficient. The best engineering teams intelligently route work across models to optimize both performance and cost.

The true AI moat: General-purpose AI hallucinates and wastes tokens. Sustainable AI advantage comes from combining models with clean, proprietary data that provides narrow, high-quality context. Increasingly, data is the moat.