So that's 1,000 cars per week per line, x 4 lines = 4,000 cars per week for 200,000 cars per year.
or, if you prefer...
2,000 cars per week per line, x 5 lines = 10,000 cars per week for 500,000 cars per year.
Figuring 2 x 10-hour shifts per line, that's 1,000 cars per shift per week per line... Ignoring for the moment that it would make sense to split the 500,000 per year across 3 different geographically-optimized plants (1 in Europe, 1 in China), meaning that each plant would only have to produce some number approximating a mere 167,000 per year...
Gotta hand it to the NUMMI collaboration between GM and Toyota that produced 500,000 cars/year almost 20 years ago now - especially given the recalcitrance of the union that killed the point of what was supposed to be GM's flagship plant in the first place.
Credit also to Henry Ford, who did what everyone said he couldn't over 100 years ago.
With the advanced robotics of today, combined with Musk's stated focus upon improving the machines that build the machines, I like Tesla's chances. And, instead of the inevitable fixation that the analysts will have with the exact numbers, what will be interesting to watch is the reduction in defect rate rather than total gross production. Once they get the defects below a certain threshold, supply chain concerns aside (small details, those - heh), they'll just ramp and ramp and ramp - rinse and repeat, if you will.