I study the origin and pricing of climate policy uncertainty. Using tools from natural language processing, I construct a novel dataset of timestamped climate policy announcements. Analyzing high-frequency returns around these announcements, I find a significant climate policy risk premium. This premium is larger when political constraints are more lax. To interpret these results, I build a model combining political economy and climate finance. In the model, climate policy uncertainty emerges endogenously because governments have private information about their future policies; this uncertainty generates a risk premium because climate policies affect cash flows and aggregate output. Political constraints affect the magnitude of the premium both by preventing implementation of extreme policies and altering the informativeness of policy communication.