Let me debunk shorts&bears thesis on "too much Model 3 cars" on the Lathrop lot.
I use to work as a young engineer in the automotive industry at both, the supply car parts company and in car manufacturer company.
Let's try to calculate the minimum number of Model 3 cars at Tesla lot and at transit for production of 5k / per week. To simplify the calculation, let assume that demand is uniform for all colors and Tesla produces only one color during a day, assembling all version that is ordered for that color. It is clear that number needed is about 60% of weekly production at lot at any moment, to be possible to send needed specification of cars versions. Practically every day the lot run out of an color. Additionally, especially during the constant rump of production, for a lot of reasons as short (love that word) supply of parts, additional QA , inconsistent shade of color etc..., lets add 700 cars more. So minimal number of cars at lot should be about 60% of 5000 plus 700 =3700.
Let's consider delivering Model 3 to TX, without considering delivery to another states which don`t allow Tesla to sell directly . Reportedly is about 500-600 contract on hold, so add some 800 cars at lot only for Texas, which increase number of cars at lot to 4500. I assume that 20 % cars is delivered directly from Fremont factory. Also, delivery by train to East coast (assuming 40 % of total production) requires another smaller lot at railway station of about 600 cars (40 wagons x 15 cars).
The numbers of cars at every lot :
Lathrop lot 3600
Fremont lot 900
Railway lot 600
Total at lots: 5100 cars
The minimum number of cars in transport for truck delivering is 3 days x 700 cars = 2100 and for railway transportation should be 8 days x 40% x 700 cars= cca 2200.
Total in transportation 4300 car.
Total cars in transit should be about 5100+4300=9400 cars at very lower end.
Anything under 10K Model 3 in transit at the end of Q3 would be great success for Tesla.