The coronavirus may have been silently spreading in the US back in September 2019, an analysis of a preprint database from the Chinese Academy of Sciences (CAS) suggests. 

Using a mathematic model and artificial intelligence, ChinaXiv, the preprint platform developed by the CAS, conducted quantitative and qualitative analysis of the infectious disease. 

By selecting 12 representative regions in the US for analysis, the date of the first infection is mostly between August and October 2019, which is earlier than the official date of the first confirmed case in the US on January 20, 2020. 

The experimental results indicate that the COVID-19 pandemic in the US started to spread around September 2019 with a probability of over 50 percent. 

Similarly, a study of more than 24,000 samples taken for a National Institutes of Health (NIH) research program in the US, between January 2 and March 18, 2020, suggested in June that seven people in five states - Illinois, Massachusetts, Mississippi, Pennsylvania and Wisconsin - may have been infected well before the country's first confirmed cases reported on January 21, 2020. 

In addition, the ChinaXiv study shows that the existing confirmed cases were also used in Wuhan City and Zhejiang Province in China to infer the time COVID-19 originated and provide results with confidence. The results show that the spread of COVID-19 in China likely began in late December 2019.