Should this statistic be sourced from Tesla, then it is likely an alternative fact (aka PR BS). Firstly let see what Musk/Tesla is comparing here: a brand new vehicles +/- 3 years all, serviced regularly (well, not really, I am hearing quite often, that Tesla does not require any maintenance at all. Excluding those unfortunate owners who’s car sits in the service centre permanently) with entire US fleet. That count all sort of cars, including those 16 years +, where I would bet is most of those fires. How about to compare Model S statistic with let say group of Mercedes S Class not older than 3 years and adjust this group to its size. I do not have data from Mercedes handy, but I would speculate, that such comparison would not ride well for Tesla, in fact it would paint quite alarming picture (assuming that the referential Mercedes group would likely have zero incidents).

IMHO, Tesla problem with fire is, that in order to drive the cost of MNC cells down, Musk is reducing amount of Cobal, which makes those cells more unstable and prone to the thermal runaway. Put that into a combination with “fast charging”, “ludicrous mode” and overall quality of those cars  …