Looking ahead to market size, I've found myself thinking more about batteries of late. I found this article:
2018 Class 8 Truck Update
Which sizes North America Class 8 truck sales somewhere between 330k and 500k (with the analysts that get paid to provide industry prognostications settling in the 330-350k range).
That might be a lot to the high side - here's another view, that also includes history, that adds in class 3-7, while also being US only:
Truck sales in the U.S. - Class 3-8 2017 | Statistic
This source is at 192k class 8 trucks in 2017 for the US, and overall 3-8 units in the 700k range.
Anyway - focus in on on the class 8's. At ~2 kWh per mile and 500 miles of range, that sounds like a 1 MWh battery pack in the Tesla Semi. It's at least a convenient number to do easy math on. For every 1k of these you want to ship per year, you need 1 GWh of battery cell and battery pack manufacturing capacity (1 MWh * 1000). I like easy math.
If you want to ship 192k of these per year, you need 192 GWh worth of GF output.
As an approximate and order of magnitude point of comparison, 500k cars/year at 80 kWh battery packs each (my guess at a blended average between the various sizes actually being shipped) comes to 40 GWh - I think I've got my zeroes in the right places. I figure this is about the rate at which packs are being built now, with a fair bit of the cells coming from Japan as 18650's, so cell output levels are lower.
The point being, this semi thing that Tesla is starting to poke at, using only the US as the target market, can absorb all of GF1 output as it exists right now, plus the next roughly 3 or 4 increments of output if Tesla were to stop shipping cars or doing anything else with the output (obviously not going to do that).
If the product is as appealing as
@jhm's been pointing out, whatever full production is for the factory, demand is there for the factory to ramp all the way to ~250 GWh/year (assuming car sales are flattish from here).
Well - I suppose Tesla won't take ALL of the market for itself, but somewhere in the world, there's going to need to be 200 GWh worth of cell and pack manufacturing going on for the US electric class 8 truck market. And that's assuming that when you build a better truck that is cheaper to operate, that your total market stays the same size - it just replaces the old market based on diesel, and doesn't grow larger.
My own guess is that road haulage in the US is going to grow. Maybe by a lot.
Oh - and I don't have pack size estimates for the class 3 - 7 commercial truck markets, but that second source suggests that's another 400k units/year. Double units and half pack size seems grossly simplistic, while double units and quarter size packs sounds anemic. Either way, I got nothing but a guess here - all I'm certain of is it's big relative to the class 8 market, and the class 8 market is huge compared to all of the EV market that has gone before, everywhere.
I'm ready to hear about construction at GF1 starting up again.