I work for one of the major players in a similar space, and the best advice I can give you is to stay away from delivery because of the capital expenditures (lots of fundraising or massive credit lines) it would take to set up that network.

The "easy MVP" of delivery will be tempting - drive the vehicle to the customer, drop it off, the employee takes uber back but I would suggest staying away from that temptation.

I'd be interested in the deprecation curves you're modeling on the vehicles as I'm just trying to ballpark what that break-even point is for the dealer. The other issue for the dealer is that while they have excess inventory, they typically have excess inventory on vehicles that aren't moving quick enough so I don't think the target is luxury models but more functional users (e.g. your situation in NYC makes sense. You need a car to get away to a destination and the car itself doesn't matter much as long as it's clean and runs)

Looking forward to seeing how you guys grow.