It is often assumed that when stars reach their Eddington limit, strong outflows are initiated, and that this happens only for extreme stellar masses. I discuss here that not even in models of up to $500mso$, the Eddington limit is reached at the stellar surface. Instead, I will argue that the Eddington limit is reached inside the stellar envelope, in hydrogen-rich stars above $sim 1 … 30mso$ and in Wolf-Rayet stars above $sim 7mso$, with drastic effects for their struture and stability. I will discuss the observable consequences and present observational evidence.