Speaking in terms of exponential growth and ignoring any negative feedback terms, the
case doubling time for the world ex-Chine is
4 days.
The current ex-China count as of Sat Mar 7th is 25000 so starting with that, we can predict the future...
caseload ~ 25000*exp(daysfromtoday*ln(2)/4) or 1 week = 84k, 2 weeks = 282k, 4 weeks = 3.2M, 6 weeks = 36M, 8 weeks = 409M, ... with the world population maxing out at T+73 days from now.
Should be noted that the doubling time in China is 21 days on account of having more negative feedback terms in the system at this point (quarantine, longer history of testing). Thus expect the doubling time to increase over time. This curve would thus be an upper limit.
Insofar the future is trending towards China and nothing gets better than the Chinese measures are at this point, then we can take
T_2=21 days as a lower limit, which looks less scary! In that case caseload ~ 105000*exp(daysfromtoday*ln(2)/21) ... where it will take 68 days to hit the first million and about 11 months to reach the world population.
However, presumably the Chinese doubling time numbers are still improving? If so, the spread could slow even more. However, with vaccines not expected for the next 12-18 months, this is definitely taking it right to the edge.