This paper reviews evidence that energy technologies and systems adapt over time to accomodate external pressures: that technical innovation and systemic change in the energy sector is largely induced by need, and restrained by potentially large transitional costs. A simple integrated model of optimal greenhouse gas abatement over time is presented, in which the abatement cost depends on both fixed and transitional elements. It is shown that the optimal current response and long-run prospects differ radically between the classical economic case - in which the cost of a given cutback in emissions is fixed exogenously - and the adaptative case - in which the response is ultimately adaptative but heavily constrained by inertia (i.e. low fixed but high transitional costs). If energy systems are indeed to a large degree adaptive, the results demonstrate that as compared with the classical non-adaptive case: long-run stabilization of atmospheric CO2 may be optimal even with moderate damages from climate change; greater near-term abatement efforts are justified; and the cost of a given delay in response may be several times higher. Neglect of the issue of induced technical change and other adaptive responses may invalidate the policy implications drawn from most integrated assessment models developed to date.