Own, Lease, or Rent the Robot

Abstract

Every capital structure decision in robotics is downstream of a single missing number: what the asset will be worth when you are done with it. Mature industrial arms carry decades of auction history behind their residual assumptions; for humanoids and the broader embodied-AI class there are zero verified secondary-market transactions — the residual is not merely uncertain, it is unpriced, and vendor-funded price deflation (a 72% decline in average selling price in two years, per one leading manufacturer's own prospectus, with margins expanding) threatens to pin early cohorts near the parts floor. The central argument is that Robotics-as-a-Service is not a financing convenience but an insurance product: decomposing a representative monthly rate puts the residual-risk premium near a fifth of the payment — sold by an underwriter who cannot price the risk either. The analysis models the buy-versus-subscribe crossover at roughly 20–30 months, weighs the tax asymmetry of permanent first-year expensing, treats vendor survival as the counterparty risk the sector's own corporate history says it is, and closes with a staged capital structure that rents the uncertainty and owns the proven duty cycle. The third paper in the robotics deployment series: REA·004 told you what to buy, REA·005 told you how not to lose the money, and this tells you how to structure the capital.