Do investors correctly price extreme events that they have never seen occur? To shed light on this question, I examine market responses to the risk of nuclear war during the Cuban Missile Crisis. I find evidence that investors indeed priced firms’ exposures to nuclear destruction: Firms headquartered in areas that American national-security experts and the general public perceived more at risk of nuclear destruction experienced lower returns. Such discrimination is plausible given contemporary survey evidence that investors generally believed that the US could recover from a nuclear war. Employing a calibrated model to reconcile survey expectations with aggregate market responses, I find that i.) investors underreacted to the risk of nuclear war; ii.) investors exhibited a far lower level of risk aversion than is standard in the literature; or iii.) investor heterogeneity or noise makes survey data inaccurate indicators of investors’ perceived exposures to extreme risks.