Context. Massive stars lose a large fraction of their mass to radiation-driven winds throughout their entire life. These outflows impact both the life and death of these stars and their surroundings. Aims. Theoretical mass-loss rates of hot, massive stars are derived to be used in applications such as stellar evolution. The behaviour of these rates in the OB-star regime is analysed, and their effects on massive-star evolution predictions studied. Methods. Dynamically consistent models are computed by solving the spherically symmetric, steady-state equation-of-motion for a large grid of hot, massive stars with different metallicities. The radiative acceleration is derived from non-local thermodynamic equilibrium radiative transfer in the co-moving frame, and all models cover a large spatial range from deep subsonic atmospheric layers into the radiation-driven and highly supersonic wind outflow. The resulting mass-loss rates are used to derive a simple scaling recipe with stellar parameters (luminosity, mass, effective temperature, and metallicity), and the new recipe is used to evaluate some first impacts upon massive-star evolution tracks. Results. We provide a new prescription for steady-state, radiation-driven mass-loss from hot, massive stars depending on their fundamental parameters. In accordance with our previous work, the rates for O stars are lower a factor of similar to 3 than the rates typically used in previous stellar-evolution calculations, where differences generally decrease with increasing luminosity and temperature. For cooler B giants/supergiants we find larger discrepancies, of up to one or even two orders of magnitude. This arises because we do not find any systematic increase in mass-loss rates below the so-called bi-stability region; indeed, our results do not show any sign of a significant bi-stability jump within the parameter range covered by the grid (T-eff >= 15 kK). Due to the lower mass-loss rates we find that massive-star envelopes are not easily stripped by means of standard steady-state winds, making it difficult to create classical Wolf-Rayet stars via this channel. Moreover, since the stars retain more mass right before they die as supernovae, our new rates make it possible to create black holes of higher masses than in previous models, even at Galactic metallicity. However, a remaining key uncertainty regarding these predictions concerns unsteady mass-loss for very high-luminosity stars close to the Eddington limit as well as the impact of non-line-driven winds.
