Businesses view a good domain name as a key part of their Internet presence. Short, memorable strings in com and net can sell for thousands or millions of dollars, or may simply be unavailable. To address this problem, ICANN has introduced alternative top-level domains (TLDs): first a few at a time, such as biz and info, but then hundreds per year starting in 2013. Alternative TLDs give registrants more opportunities to get a name they will like, but give the same chances to name speculators. Additionally, existing companies find themselves with new namespaces in which to defend their trademarks. One would hope most registrants want domain names to develop an Internet presence, and that defensive registrations and name speculation are the exception. Unfortunately, few studies quantify domain name registration types, and none of them at scale. This dissertation identifies the intent of registrants in hundreds of TLDs. We combine registration intent with user visit data and pricing information to provide a comprehensive view of each TLD in our set. We open the dissertation with two case studies, first of biz and then of xxx, and quantify the types of registration behavior in each. Then we combine the lessons learned from each to scale our methodology to hundreds of TLDs in ICANN's New gTLD Program. We find that defensive and speculative domain registrations are extremely common, and that some TLDs seem to encourage them. Domain registrants in xxx spent $24 million USD in the first year to defend their names and trademarks, and less than 7% of yearly fees come from primary registrants. By contrast, over 30% of new domain registrations in com host legitimate content. ICANN's New gTLD Program generated new domain registrations, instead of simply shifting them from old TLDs to new ones, but most of them receive low quality registrations. Parked domains make up 32% of registrations in the new TLDs, and free promotional domains cover 12% of the space. Even our most permissive pricing models show that roughly 10% of registries will never become profitable at current registration rates