Serve earns a better multiple only if investors can see a real operating system for recurring physical tasks, not a single-channel delivery experiment. I assume some
rerating from delivery recovery, hospital contracts, and workflow verification, but I cap it well below premium software because the model should remain service-heavy, capital-aware, and somewhat price sensitive.