We have recently reported multiple lines of evidence in favor of a stochastic background of gravitational waves in the NANOGrav 15-yr dataset. This result is consistent with findings from other collaborations around the world. In this talk I will discuss the dataset, analyses, and interpretation of our findings, as well as future work. The latter includes a detailed spectral characterization of the signal we have found to better understand its origin, and searches for individual signals that will enable multi-messenger observations of the largest black holes in the universe.