OK, bonaire, here's the problem: you seem to be running an accumulating count of cars going into the inventory / loaner / internal use fleet, which seems to be sound, but you don't appear to be properly calculating cars coming *out* of that fleet. Some of them are sold as new, but some of them are being sold as used, not as new; we have evidence of this. So you're slowly getting an inflated cumulative estimate. (Any car going to crash testing will also be produced and not delivered and will inflate your estimate. And so on.) I don't think you can add up numbers over time in the way you are doing without ending up with an overestimate of the inventory.
If a car goes into fleet for loaner, it can be sold as new in almost every location after months of use. Other than Colorado which says any "new" car must have less than 1500 miles. I know there has to be some loss in the system. Crash tests, test-drive crashes (sometimes the customer is on the hook to pay for it) and other factors. It's a lossy system and the 16500 total produced but not sold will have some overall losses. Doesn't factor in the overage of Vin #s as well for those gaps that happen over time. How many crash test cars will there be per year? 20? 10? How many will get crashed worldwide during test-drives? hundreds maybe? Europe had 1-week test drives allowed - I suppose customers are responsible to pay if they put one in a ditch during their test drives and it is written off as totaled - I suppose that is a sale (to the insurer) rather than an inventory write-down.
The only way to sell them as used is to self-title the cars (ie. Hong Kong, Denmark, Norway?) and then sell as used with a lower price and capture the incentive internally. I think any "new" car that has more than 10,000 miles on it should be considered used and not qualify for the regional incentives of the country/state/municipality. I am not a CPA but isn't a firm able to write the miles used to sell a car down as a "Cost of Revenue"? And if done using the IRS formula in the USA, that is $0.54/mile? Though, as a loaner - it is more of a cost-of-servicing a customer, something like that.
What is the industry average of fleet-losses prior to sale? I hope it isn't very much. That's expensive write-downs. If we can determine a fleet-wide auto industry scrapping loss of actual production cars, we can figure something out. I go with maybe 1000 units per year - so maybe 4000 since inception, taking 16500 down to 12500. That's a lot of cars - and amounts to $400 Million in possible lost-sales with an ASP of $100k each. If they are that loose with scrapping - why was there such a furor over discounting in Q3 of 2016 for a few k off a car to make a sale? Expense management also included trying to ship by train to east coast buyers rather than truck to save a few bucks. I suspect scrapping is far less than 1000 units a year. But who would really know? Q1 Inventory Writedown was: $26,918,000 - pretty hefty. Inventory is well known as "finished goods" - cars, powerpacks, powerwalls, solar stuff. But what does that quarter's inventory writedown include? 2016 year-long inventory write-down was only $65,520,000. So, what happens behind the curtain is hard to know.
Now, if I see an inventory car sitting in Fremont and I want it shipped to me, I must pay $2k extra. And it hasn't even left Fremont yet. Expense management seems tight there.
Look my numbers over here and I had to take guesses on the MS production the last three quarters because they stopped breaking them out.
