# Capital Strategy in a Hype Cycle

**Track:** Scaling & Stewardship — AI for Entrepreneurship — complete (29)
**Framework / surface:** venture strategy
**Level:** Advanced
**Prerequisites:** Unit Economics of Inference

**In one line:** Raising when everything is overpriced: milestones investors believe, and burn discipline the curve rewards.

## Theory, aesthetics & inspiration

Carlota Perez's "Technological Revolutions and Financial Capital" (2002) reads every great technology surge the same way: frenzied capital overshoots, crashes, and then the durable deployment period builds the real economy — a rhythm Gartner's hype cycle restates in miniature. Raising in the frenzy phase is a discipline of its own: capital is abundant but priced on narrative, and the valuation that flatters today becomes the bar that indicts tomorrow. The AI-specific traps are twofold — burn that scales with usage (inference is a variable cost venture math often models as fixed) and milestones borrowed from the platform story ("we'll train our own model") rather than the business ("we'll own this workflow"). David Sacks's burn multiple — dollars burned per dollar of net new recurring revenue — travels well here, with one AI-native amendment: report margin per outcome alongside growth, because investors have learned that AI revenue can be services in a software costume. Raise against evidence the curve will strengthen: retention, data loops, and evals — not against a demo the next release gives everyone.

**Founder question:** What milestone would you still be proud of if the hype cycle ended tomorrow?