On the premise that knowledge creation defines contemporary metropolitan regions, we profile them by their inventive networks, as measured by a variety of complementary social network, technology, and patenting metrics that distinguish scalar and structural aspects. Using a comprehensive, multiyear database of patent applications, we investigate whether the knowledge creation network profiles are discriminating characteristics of metropolitan regions by establishing a new urban taxonomy for metropolitan areas in the United States. The four-class taxonomy is not only statistically significant, but it is also economically meaningful in terms of economic performance of metropolitan areas. We find that metropolitan areas benefit from a higher density of inventors in the population, and that there is a positive correlation between economic performance and metropolitan areas with inventor teams working in similar or complementary areas of technology. In fact, the structure of knowledge creation networks are fundamental to economic performance and extends to metropolitan growth rates in jobs and income.