This study examines the bushfire (wildland fire) risk to the built environment in Australia. The most salient result is that the annual probability of building destruction has remained almost constant over the last century despite large demographic and social changes as well as improvements in fire fighting technique and resources. Most historical losses have taken place in a few extreme fires which if repeated are likely to overwhelm even the most professional of fire services. We also calculate the average annual probability of a random home on the urban–bushland interface being destroyed by a bushfire to be of the order of 1 in 6500, a factor 6.5 times lower than the ignition probability of a structural house fire. Thus on average and if this risk was perceived rationally, the incentive for individual homeowners to mitigate and reduce the bushfire danger even further is low. This being the case and despite predictions of an increasing likelihood of conditions favouring bushfires under global climate change, we suspect that building losses due to bushfires are unlikely to alter materially in the near future.