I examine the impact of patent term on R&D and innovation in the presence of policy anticipation, common in real-world settings. Using a difference-in-difference design, I exploit quasi-experimental variation in U.S. patent term across technological fields due to the ratification of TRIPs agreements in 1995. Despite a general increase in average patent term, in most fields innovators faced a considerable probability of patent term reduction for future innovations. Three key findings emerge: (1) R&D and innovation accelerate more in fields with a higher probability of patent term reduction, i.e., a shorter average patent term extension, before implementation. (2) This heightened activity persists for at least five years post-implementation, driven by indirect effects where the news-related acceleration fosters further innovation through technological externalities linked to cumulative knowledge creation. (3) Conversely, the direct effect of a shorter extension in patent term would stimulate relatively less innovation, absent the indirect effects of anticipation.