In early-age concrete, expansion due to temperature increase followed by contraction due to temperature decrease and shrinkage are produced by the hydration reaction. If restrained, these volume changes generate compressive stresses followed by tensile stresses that may reach the concrete tensile strength, resulting in cracking. This stress history causes concrete creep, which either increases or reduces the cracking risk depending on whether creep dominates in the heating or cooling phase. Creep is additionally influenced by the hydration-induced temperature variation of the concrete, particularly significant in massive structures. The characterization of creep at early ages for massive restrained concrete structures is hence of importance in cracking risk analysis. The current paper shows the results of an experimental investigation on concrete creep at early ages performed in the Temperature-Stress Testing Machine (TSTM) at NTNU. Creep tests under compression in the heating phase, tension in the cooling phase, and compression and tension applied subsequently in the two phases were performed. The tests were conducted under isothermal conditions and two simplified versions of the realistic hydration-induced temperature history of the concrete in a thick wall. The use of the same equipment for both compressive and tensile creep tests, allowed to directly compare the test results with minimum uncertainty, showing the validity of the principle of superposition under all temperature conditions. The results show that the instantaneous and delayed effects of the temperature variation on both compressive and tensile creep, could be adequately described by the transitional thermal creep theory and the maturity principle.