By 11 May, when lockdown restrictions were eased in France, about 4.4% of the French population had been infected with SARS-CoV-2, a new modeling study suggests. These estimates are well short of what would be required for herd immunity, say the authors. Understanding the level of immunity to SARS-CoV-2 will be key to avoiding a rebound in the epidemic as populations around the world ease lockdowns. Serostudies - which detect anti-SARS-CoV-2 antibodies in patients' blood - are important to informing questions of immunity, but growing evidence suggests that many people, especially those who are asymptomatic, take a while to seroconvert. During this time, researchers seeking to understand virus immunity need to rely on indirect measures, including data on hospitalizations and deaths. France has been heavily affected by the SARS-CoV-2 epidemic and went into lockdown on March 17, 2020. Following the expected reduction in cases, the French government announced it would ease restrictions on 11 May. To better understand the underlying level of population immunity and infection at this time, Henrik Salje and colleagues studied de-identified hospital records of COVID-19 patients from all hospitals in France, along with surveillance data from these hospitals and from the Diamond Princess cruise ship. About 3.6% of infected individuals in France were hospitalized and 0.7% of those infected died, they say, with 10.1% of infected people over 80 dying. Across all ages, the authors report, men were more likely to be hospitalized than women. Applying their data to national models of disease transmission, Salje et al. estimate the lockdown in France resulted in a 78% reduction in transmission. By 11 May, when restrictions were eased in France, 4.4% of the population was likely infected, though infection rates were closer to 10% in two French regions most affected. However, because around 65% of the population would need to be immune for the epidemic to be controlled by immunity alone, population immunity appears insufficient to avoid a second wave, say the authors. Efficient control measures need to be maintained. The authors note their model projections can support healthcare planning of the French government by forecasting hospital bed capacity requirements.