We compared the microclimate experienced by maternity groups of eastern cave bats (Vespadelus troughtoni) in northern New South Wales between roosts in overhang caves versus a nearby maternity roost beneath the corrugated iron roof of a farm shed. The shed’s microclimate differed dramatically from that of the caves. Caves provided well-buffered microclimate conditions, which were cooler and more humid than ambient conditions during the day, but were warmer and less humid than ambient at night. Early summer temperatures remained between 20 and 25°C in the maternity caves. In contrast, the microclimate at the shed was not buffered, with conditions varying enormously over a 24-h period, being hotter than ambient during the day (mean = 30°C, maximum = 40°C in the middle of the day), and similar to ambient at night (mean = 18°C at dawn). Rather than selecting thermal stability, lactating females selected warm, but variable, diurnal temperatures in the shed, presumably to sustain lactation. Under this situation, we estimated considerable energetic savings by roosting in the shed (27–35%, assuming normothermia) compared with that for caves, and even higher savings if bats were torpid in the mornings and passively rewarmed later (60–69%). Lactating bats therefore presumably traded-off the cost of leaving juveniles behind in a cool roost at night with these energetic benefits. In late autumn bats continued to roost in the shed with day temperatures rising above 30°C, compared with unoccupied caves, which were considerably cooler at this time (~20°C). Clarification of the estimated energy savings and the possible costs in terms of milk production requires further research to measure concurrently body temperatures, roost temperature and energy expenditure, together with juvenile growth in different types of roosts.