Uncertainties regarding the axial cyclic behaviour of piles driven in sands led to an extended programme of calibration chamber instrumented pile experiments. Broad trends are identified and interpreted with reference to normalised cyclic loading parameters Qcyclic/QT, Qmean/QT and N. Cyclic damage is shown to be related to changes in the radial effective stress regime close to the shaft. While stable loading leads to little or no change as cycling continues in the sand masses’ effective stress regime, high-level cyclic loading can affect stresses far out into the sand mass. The test systems’ chamber-to-pile diameter ratio has a significant impact on outcomes. Piles installed in loose, fine, sand are far more susceptible to cyclic loading than in denser, coarser sand. Little or no change in pile stiffness was seen in tests that remained within the stable cyclic region, even over 10,000 or more cycles. Unstable tests lost their stiffness rapidly and metastable cases showed intermediate behaviours. The permanent deflections developed under cycling depend on the combined influence of Qcyclic/QT, Qmean/QT and N. While model tests provide many valuable insights into the behaviour of piles driven in sand, they are unable to capture some key features observed in the field.