Historical evidence suggests natural disturbances could allow more forest persistence, than expected from models, over 40 yr of transition to the net‐zero emissions needed to limit warming to <2.0°C (e.g., Paris Agreement). Forests must ultimately equilibrate with committed warming from accumulated emissions. Historical dry‐forest landscapes were heterogeneous from large, infrequent disturbances (LIDs) that reduced tree density and basal area, followed by slow, variable tree regeneration and recovery for 1–3 centuries. These together effectively provided bet‐hedging through stand‐ and landscape‐level heterogeneity that enhanced resistance and resilience to a diversity of unpredictable subsequent disturbances. Recent disturbances have not yet exceeded historical variability in rates and patterns, but could cause mortality of ∼26–51% of dry‐forest area in the transition. This also means 1/2 to 3/4 of dry‐forest area could escape most mortality and the mortality area could also have substantial forest persistence. Projections are unavailable for droughts or beetle outbreaks, but they recently caused about 3–4 times as much tree mortality as did moderate‐ to high‐severity fires. Mortality could reduce forest area if new trees do not regenerate, but 24 studies showed recent regeneration after high‐severity fires was slow, but indistinct from historical variability. Survival of smaller trees provided regeneration after beetle outbreaks and droughts. Regeneration in general was projected by 2060 to decline by ∼10% in one study and increase by 50% in another. If openings from disturbances increased, some grasslands and shrublands could be restored, increasing landscape heterogeneity and resistance to disturbance spread. Given these trends and our limited ability to prevent LIDs, I suggest (1) refocusing restoration to increase bet‐hedging resilience to droughts and beetle outbreaks by retaining small trees and diverse tree species, (2) expanding development of fire‐safe landscapes to protect people and infrastructure from unavoidable increased fire, (3) enabling more managed fire to restore and enhance stand‐ and landscape‐scale bet‐hedging, and (4) accepting that LIDs will revise resistance, resilience, and adaptation, which enhance forest persistence, particularly if post‐disturbance survivors are not logged and trees are not planted. Natural disturbance and slow recovery, if bet‐hedged to increase resistance and resilience, could enable substantial forest persistence.