This paper analyzes human capital externalities from high-skilled workers by applying functional regression to precise geocoded register data. Functional regression enables us to describe the concentration of high-skilled workers around workplaces as continuous curves and to efficiently estimate a spillover function determined by distance. Furthermore, our rich panel data allow us to address the sorting of workers and disentangle human capital externalities from supply effects by using an extensive set of time-varying fixed effects. Our estimates reveal that human capital externalities attenuate with increasing distance and disappear after 25 km. Externalities from the immediate neighborhood of an establishment are twice as large as externalities from surroundings 10 km away.