Whether that's likely or not can be estimated by extrapolating case numbers from January back in time into 2019 and using a reasonable estimate for the R0 value without any measures against the spreading of the disease. Even if you don't trust the Chinese numbers, the estimate should be roughly in the right ballpark, since we certainly didn't have millions of deaths in Wuhan in January. The initial exponential growth phase is similar in all regions with comparable population density.

I haven't crunched these numbers, but a start in early September seems unlikely to me - call it a hunch, based on looking at the initial growth curves.