Could be, that the first 100,000+ cars will bring less margin. If they've ironed out the teething problems well enough before non-employees get theirs (slow ramp up helps here), the overhead cost may be below budget. For haf a million cars per year, a year from now.
Soooo.... Model S is needlessly heavy? ;-)
Well, perhaps Model 3 is indeed much lighter after all, and not just to be contributed to the 2170 cells with staggering density.
I feel that Model S still has a good number of years to go, but it may be in for an under-the-skin full redesign, taking lessons from Model 3 design and also incorporating the improvements being worked on for Model Y (much less cabling). If they redesign Model S to be easier (cheaper) to produce, they can boost its longuevity in an increasingly competitive market through cost reductions and remaining "current". Today there's no competitition, but within 3 years that'll be different. They'll need to offer bigger packs, more versatility and keep those cars out off service visits.
Imagine Lucid come through with a $45K version of that Air car, 2019-2020'ish. Model 3 doesn't compete with that. Likely Model Y will be late or underspec'd. Model S will need to step it up.