Here's an idea.... Tesla needs to make a profit and has determined that paint is their current bottleneck. In order to be profitable they have eliminated (temporarily probably) their two least taken paint colors.
My thought process is they know better than us how much this will improve their production and delivery numbers for the remainder of 2018.
I'm not claiming anything regarding the reduction of color options or feasibility, nor critiquing Tesla's choice. Only stating that @arnis's suggestion does not require 10s of thousands of cars sitting around for months.
1: Collect orange orders for 2 months (or until threshold of profitability is reached)
2: Build that number of orange cars
3: Ship orange cars. I'm guessing there was a slight translation issue in typing "as soon as all oranges are ready" the 'all' was not meant/needed as a criteria of shipping any of them.