Assuming that dark energy may be treated as a fluid with a well defined temperature, close to equilibrium, we argue that if nowadays there is a transfer of energy between dark energy and dark matter, it must be such that the latter gains energy from the former and not the other way around.