Background The coronavirus disease (COVID-19) pandemic began in Wuhan, China, in December 2019. Wuhan had a much higher mortality rate than the rest of China. However, a large number of asymptomatic infections in Wuhan may have never been diagnosed, contributing to an overestimated mortality rate. Objective This study aims to obtain an accurate estimate of infections in Wuhan using internet data. Methods In this study, we performed a combined analysis of the infection rate among evacuated foreign citizens to estimate the infection rate in Wuhan in late January and early February. Results Based on our analysis, the combined infection rate of the foreign evacuees was 0.013 (95% CI 0.008-0.022). Therefore, we estimate the number of infected people in Wuhan to be 143,000 (range 88,000-242,000), which is significantly higher than previous estimates. Our study indicates that a large number of infections in Wuhan were not diagnosed, which has resulted in an overestimated case fatality rate. Conclusions Increased awareness of the original infection rate of Wuhan is critical for proper public health measures at all levels, as well as to eliminate panic caused by overestimated mortality rates that may bias health policy actions by the authorities.