Abstract Looking ahead to 2050 many countries intend to utilise wind as a prominent energy source. Predicting a realistic maximum yield of onshore and offshore wind will play a key role in establishing what technology mix can be achieved, specifying investment needs and designing policy. Historically, studies of wind resources have however differed in their incorporation of physical limits, land availability and economic constraints, resulting in a wide range of harvesting potentials. To obtain a more reliable estimate, physical and economic limits must be taken into account. We use a grid-cell approach to assess the theoretical wind potential in all geographic locations by considering technological and land-use constraints. An analysis is then performed where the Energy Return on Investment (EROI) of the wind potential is evaluated. Finally, a top-down limitation on kinetic energy available in the atmospheric boundary layer is imposed. With these constraints wind farm designs are optimized in order to maximize the net energy flux. We find that the global wind potential is substantially lower than previously established when both physical limits and a high cut-off EROI > 10 is applied. Several countries’ potentials are below what is needed according to 100% renewable energy studies.