Off the top of my head, wherever the marginal returns from an an increase in ecficacy are very large, either because the outcome is very sensitive to some factor or the prize is just very big. So places with strong feedback loops where you can consolidate the gains, and you know others are doing so as well. Thus, probably just about any technology that is suddenly in demand due to an unexpected event, for one. Another may be any sort of strong truly defensive (protective, not deterrent) capability. Obvious yet profitable applications of new technology fall into the event category above, if first to market wins most of market share.

May also be worth thinking of anti-moloch factors, resulting in underinvestment relative to gains. Market share will be split many ways, first mover has little advantage, and perhaps greater costs. So maybe lack of strong IP protection. What else?