The era prior to 1990 was a time of careful observation of disease presentation, course, outcomes and meticulous pathology studies. These mainly single-centre studies introduced new life-saving therapies for drugs still used effectively today. In the 1970–1980s, cyclophosphamide (CyP) added to glucocorticosteroids (GCS) was shown to be life-saving. The trade-off was often severe adverse events. Some forms of vasculitis were found not as ominous as thought initially. Some could be treated with safer drugs [e.g. methotrexate (MTX)]. However, whether mild or severe, patients were not cured. From 1990 to the present large collaborative networks have provided studies were not possible heretofore. Randomized controlled trials captured and manipulated vast amounts of data, banked biological specimens and shared these resources and intellectual capital, moving the field forward at an extraordinary pace. We now know that even for severe forms of granulomatosis and polyangiitis [granulomatosis with polyangiitis (GPA), Wegener's granulomatosus (WG)], microscopic polyangiitis (MPA) and Churg–Strauss syndrome (CSS), we do not need to use CyP for extended periods. We have learned recently that rituximab is as effective as CyP for severe WG and MPA. We should never again see the permanent toxicities born from years of chronic CyP use. However, short courses of CyP remain useful and can be life-saving. Step-down therapy from CyP is now a standard of care, perhaps to be replaced by rituximab in the future. If one accepts the premise that there are few cures at present for idiopathic large- and small-vessel vasculitis, we will serve our patients well if we can determine the most effective initial therapy that leads to a maintenance strategy for remission with least risk. Ultimately, we wish to identify causes of vasculitis so they can be used as a wedge to secure cures. Unmet needs and strategies are as follows: (1) to increase the numbers of vasculitis-trained physicians; (2) to define risk–benefit formulae for chronic maintenance therapy versus discontinuation of treatment after remission; (3) to define risk– and cost–benefit formulae for laboratory monitoring; (4) large-scale studies with longer follow-up that explore inhibition of interleukin-5 in CSS; (5) to explore the value of anti-interferon-γ for GCA, Takayasu's and other granulomatous vasculitides; and (6) identification of aetiological factors: cures will probably be linked to knowledge of the antigen driving the disease, plus vulnerabilities of the patient that prepare them to develop an illness phenotype. Improved outcomes using anti-inflammatory/immunosuppressive agents do not rule out infection as a driver for autoimmunity. Techniques that can facilitate pathogen discovery have never been more sophisticated.