All eyes are focused on reaching the 5K/week production speed -- which is important for cash flow and margin. However, I see a much bigger issue with the overall model 3 production ramp plans.
The problem I see is that the end-goal seems to be 10k/week, and I do not think that is going to cut it! When the first half million reserved model 3 is on the roads visible to everyone and new ones can be ordered with a reasonable wait time (3-4 months), then demand will be more than BMW 3-series + Audi A4/S4 + Mercedes C-class added together world-wide.
Tesla will need to ramp production far higher than half a million cars per year to satisfy that. Can they put 2 more M3 lines into Fremont (in addition to the 2 planned initially) ? Can the Gigafactory (once fully built out) output enough battery packs for a million M3 per year ? If not, where are they going to build the extra capacity ?
This probably doesn't apply to the 3, but: ramping production to match demand is not always the best route. Eventually demand is lower, or the product/ manufacturing line is obsoleted and then you have less utilized capital equipment. For an example, look at the Nintendo approach to the Wii, they had a huge backlog of orders, but kept manufacturing as-is.
That said, if the Y is mostly common with the 3, and the next version Dreadnaught is even more efficient, then the Y line may also support additional 3 volume (or it may be swamped with Y orders).
GF1 is only 33%? completed, so lots of extra future capacity there to support S/X/3/ Semi/ Roadster/ PW/ PP packs.
At some point, Fremont will hit the physical limit of material in/out flow. 1 million cars per year is 2,700 per day, 114 per hour, 13 car haulers per hour, or one every 4.73 minutes 24/7/365. And that is finished goods, the material input will also be immense.
Is there still a limit in terms of paint booth emissions?