Purpose: Increasing the longitudinal bending stiffness of a running shoe decreases metabolic energy cost. Optimal stiffness occurs at the stiffness with the lowest metabolic cost. However, it is still unclear how optimal stiffness varies among individuals. The primary purpose of this study was to understand how optimal shoe stiffness is affected by running speed. A secondary purpose examined the anatomical and biomechanical characteristics associated with optimal stiffness variation. Methods: Six shoe stiffness conditions were manufactured by inserting carbon fiber plates between the midsole and outsole of a standard running shoe (shoe stiffness range: 9.26–23.83 N/mm). Twenty-one experienced male runners (mass = 67.1 ± 5.0 kg, height = 178.9cm ± 4.0 cm, age = 26.4 ± 8.4 years, American shoe size 10–11.5) completed testing at a slow 2.98 m/s and fast 4.47 m/s speed over two testing days, completing 5 min in each shoe condition. Metabolic cost was measured along with several biomechanical and anatomical variables. Data were also separated by foot strike to allow additional analysis. Results: At the fast speed, average optimal stiffness was 19.29 N/mm (± 5.62) with a metabolic benefit of 3.02% (± 2.62%). Slow speed average optimal stiffness was 17.04 N/mm (± 6.09) with a metabolic benefit of 1.93% (± 1.82%). Only rearfoot strikers demonstrated a significant increase in optimal stiffness (p = .020) and the associated metabolic benefit (p = .024) across speeds. There were no correlations between any of the measured anatomical or biomechanical variables and optimal stiffness. Conclusion: Optimal stiffness varied between subjects but was not correlated to any of our measured characteristics. Rearfoot striking runners may benefit from a high stiffness shoe at faster speeds to enable optimal performance.