The dwindling supply of new antibiotics largely reflects regulatory and commercial challenges, but also a failure of discovery. In the 1990s the pharmaceutical industry abandoned its classical ways of seeking antibiotics and instead adopted a strategy that combined genomics with high-throughput screening of existing compound libraries. Too much emphasis was placed on identifying targets and molecules that bound to them, and too little emphasis was placed on the ability of these molecules to permeate bacteria, evade efflux and avoid mutational resistance; moreover, the compound libraries were systematically biased against antibiotics. The sorry result is that no antibiotic found by this strategy has yet entered clinical use and many major pharmaceutical companies have abandoned antibiotic discovery. Although a raft of start-up companies-variously financed by venture capital, charity or public money--are now finding new antibiotic compounds (some of them very promising in vitro or in early trials), their development through Phase III depends on financial commitments from large pharmaceutical companies, where the discouraging regulatory environment and the poor likely return on investment remain paramount issues.