Carbon nanotubes (CNTs) are novel materials with remarkable properties; possible beneficial applications include aircraft frames, hydrogen storage, environmental sensors, electrical transmission, and many more. At the same time, precise characterization of their potential toxicity remains elusive, in part because engineered nanostructures pose challenges to existing assays, predictive models, and dosimetry. While these obstacles are surmountable, their presence suggests that scientific uncertainty regarding the hazards of CNTs is likely to persist. Traditional U.S. policy approaches implicitly pose the question: “What level of evidence is necessary and sufficient to justify regulatory action?” In the case of CNTs, such a strategy of risk analysis is of limited immediate utility to both regulators essaying to carry out their mandates, and users of CNTs seeking to provide an appropriate level of protection to employees, customers, and other stakeholders. In contrast, the concept of anticipatory governance suggests an alternative research focus, that is: “Given the conflicted character of the data, how should relevant actors respond?” Adopting the latter theoretical framework, this article argues that currently available data support treating CNTs “as if” they are hazardous, while simultaneously highlighting some systemic uncertainties in many of the experiments carried out to date. Such a conclusion implies limiting exposure throughout product lifecycles, and also points to the possible applicability of various conceptual tools, such as life-cycle and multicriteria decision analysis approaches, in choosing appropriate courses of action in the face of prolonged uncertainty.