Nanotechnology exposes a structural weakness common to the regulation of emerging technologies: legal authorization must often be granted before the scientific properties of the regulated object – and even its proper legal classification – are fully understood. This paper develops a formal law and economics model of dynamic regulatory licensing designed for precisely this setting. Departing from conventional regulatory theory, the model treats the regulator not as a Bayesian expected-utility maximizer but as a decision-maker operating under genuine ambiguity, facing a set of plausible probability distributions over harm rather than a single, known one. Framing the licensing decision as a dynamic optimal-stopping problem, the paper characterizes the welfare-maximizing policy as one in which authorization status, monitoring intensity, and regulatory obligations are continuously recalibrated in response to an ambiguity-adjusted harm index. Two extensions refine this baseline result: incorporating realistic administrative costs of granting and revoking authorization yields an optimal licensing rule in the form of a two-threshold hysteresis band rather than a single frictionless cutoff, and, in a setting where the innovator holds private information about the technology’s risk type, the socially optimal incentive-compatible license menu requires monitoring the riskier type below its own first-best level, leaving an information rent to the safer type. The comparative analysis of European Union and United States nanotechnology governance developed in the paper shows that the European precautionary model and the American post-market model are each optimal only under conditions that nanotechnology systematically violates, and numerical simulations quantify the welfare gains of dynamic licensing relative to both benchmarks. The paper concludes by addressing the institutional implementation of the model – the allocation of regulatory functions between legislatures and specialized agencies, and the incentive-compatible production of regulatory information – and argues that the underlying framework, though developed through the empirical lens of nanotechnology, applies more broadly to any technological domain in which scientific understanding evolves throughout, rather than before, commercialization.