This analysis examines the spatial fragmentation of the urban landscape with respect to neighborhoods classified according to their racial, demographic, housing and socioeconomic characteristics. The analysis is performed on the 50 largest metropolitan areas throughout the United States from 1990–2010, and looks at both global trends over time using a landscape ecology metric of edge density to quantify fragmentation over time. It then analyzes the spatial clustering of each neighborhood type over time, for each city. Results illustrate an increasingly fragmented urban landscape with respect to neighborhood type, led by Los Angeles as the most fragmented metropolitan area. Decomposed by neighborhood type, both racially concentrated high-poverty neighborhoods, as well as neighborhoods with a highly educated population, have increased in spatial concentration in large cities over time, exposing rises in spatial inequalities even as global patterns suggest a breaking up of neighborhood types. The global patterns are therefore driven by declines in more moderate-income and multiethnic neighborhoods, and a decline in the spatial concentration of newer, white, single-family housing neighborhoods.