We study the interactions of human capital, growth and inequality by embedding networks into an endogenous growth model with overlapping generations. Human capital depends on investment in education and the average human capital of a household's neighborhood. High network cohesion leads to long run equality, while for low network cohesion inequality is high and persists more often. During transition, high overall growth is achieved when the network has high degree centralization, and high individual growth is achieved when the household has low human capital relative to its neighborhood and is located in a neighborhood with high average human capital.