 This paper demonstrates the power of combining first principles data-driven simulations with advanced enhanced sampling algorithms to accurately model the thermodynamics of water's phase diagram. The results provide new insight into the role of enthalpy, entropy, and quantum effects in determining the free energy landscape of water. This work also shows that recent advances in first principles data-driven simulations are now capable of producing accurate predictions of complex molecular systems, bridging the gap between experiment and simulation. This article was authored by Siegge and Landboer, and Francesco Paesoni.