I definitely see where you're coming from. Even though I generally agree with you that the "95%" calculation is probably not allocating expenses appropriately, I guess I'm looking at it slightly differently:
Say a sales location costs Tesla ~$300k per year to operate and it has so far sold 500 Model S/X cars per year.... Tesla still has to expense that full amount, essentially at a $600 per car rate.
When the Model 3 comes out and that sales location is now selling 5,000 cars per year with 5x the staff (so say ~500k per year to operate since some is fixed lease costs), then the rate drops to $100 per car.
GAAP does not capture this leverage benefit. It also doesn't capture the benefit of having that sales location there for two years before the Model 3 comes out, expanding the brand value for Tesla. And imagine that Tesla has already hired the 2-3x the people for training purposes six months before the Model 3 ramps up. GAAP doesn't adjust for that either; there's no capitalizing of salespeople salaries.
GAAP generally can be conservative for high-growth companies, and Tesla is a super-high-growth company....
This explanation cannot be applied to all high-growth companies, but in Tesla's case, I believe this is how the GAAP vs. non-GAAP divergence is playing out (among other things). It's all situation/company based.
To reiterate though: I generally agree with you that the "95% not applicable to S/X" comment doesn't make much sense and shouldn't be taken at face value, but I also think Tesla will see much higher SG&A leverage than most expect.
