In this Article, we empirically study the use of administrative validity challenges by defendants in patent infringement suits. By requesting an administrative challenge, defendants can effectively bifurcate a patent infringement suit: staying district court proceedings while they challenge the validity of the patent at the patent office. Because of potential advantages in cost, speed, and legal standards, administrative challenge procedures like inter partes review appear facially attractive to defendants, and have been heralded by scholars as a way to reduce litigation costs and improve the patent system. Despite all of the potential benefits, we find that defendants requested an administrative challenge–inter partes reexamination or inter partes review–in only about ten percent of the sixty thousand infringement cases brought between 2008 and 2015. Some of the low challenge rate can be explained by statutory ineligibility and changes in the standard for the joinder of multiple defendants. But most of the low challenge rate appears driven by speedy resolutions of the underlying dispute: over fifty percent of the cases where defendants did not use a challenge settled or otherwise terminated within one year. Our results have three important implications. First, we discover trends that can inform the design and evaluation of administrative challenge procedures. We find substantial evidence that the 2011 America Invents Act reform increased the use of administrative challenges by defendants. Still, the reform had heterogenous, possibly unintended effects. Small entity patents, for example, are much less likely to be challenged after the AIA than before. We also identify areas ripe for future reform. Despite a growing policy concern over patent assertion entities (PAEs), for example, patents asserted by PAEs are currently less likely to face an administrative challenge than other similar patents. Second, we empirically test several predictions in a growing theoretical literature on the interaction between Article III courts and administrative venues. We confirm an important earlier result: a defendant’s decision to request an administrative challenge appears highly sensitive to the district in which they are sued. Pushing further, we also discover sensitivity to the particular judge assigned, which is driven at least in part by each judge’s docket management. Third, we caution that a growing empirical literature on the outcomes of inter partes review may be clouded by selection bias. We find some evidence that the cases where defendants use administrative challenges involve patents of particularly broad scope and high value compared to the broader pool of litigated patents. Our selection models provide a framework for future authors to consider and account for these selection effects.