An unruptured brain arteriovenous malformation (AVM) is one found before it has bled, after a seizure or headache or on imaging done for another reason. The clinical question is whether a low but lifelong risk of haemorrhage justifies the up-front risk of treatment. The decision to treat or observe is weighed against the ARUBA randomised trial, which favoured medical management alone, and against the consensus of the European neurosurgical, neurointerventional and radiosurgical societies (EANS, ESMINT, EGKS), which frames the decision by Spetzler-Martin grade. When treatment is chosen, the size, location and grade of the lesion guide the choice between microsurgery, stereotactic radiosurgery (SRS) and embolisation. Microsurgery removes the AVM at once, whereas SRS obliterates the nidus only after a latency of years, so imaging follow-up after SRS is long.