There are ~91 days in a quarter. Weekends and holidays are included.
Ok, thanks for the chart. Here's my full feedback --
1. As already stated, I would start the chart in 2020 and assume only new cars built that year qualify.
2. 25% revenue to Tesla is likely correct, based on Elon's previous comments and other industry norms.
3. You're assuming ~150,000 miles per year per car. (7 miles per paid trip + 7 miles return at no revenue) * 2.5 trips/hour * 12 hours * 365 days. This is also probably in the ballpark.
4. However, assuming 12 hours/day every day of the year means these are
dedicated taxi cars not owners renting them out.
I do believe 3rd party taxi companies will form around the Tesla network. 2% may be about the right number for a dedicated service. HOWEVER -- you're assuming an additional 2% increase every quarter IN ADDITION TO the new cars built that quarter. So, your chart shows a geometric increase (from day 1) in the number of wholly dedicated "taxi" cars every quarter. It's not going to happen like that. Technology adoption follows an S-curve. It takes a time for people to accept something radically new, then it ramps linearly then it ultimately levels off.
Further, keep in mind that the Model 3 is backordered -- not by taxi companies but by people. This backorder will not be depleted by this time next year.
So you're projecting that 2% of regular people (mostly wealthy already), geometrically increasing every quarter, will fully dedicate their $60-100K vehicles to pick up strangers 12 hours every day of the year. Does that sound realistic?
How many automated superchargers do you think will be in place by this time next year? Enough to cover 2% of the fleet for 12 hours/day everywhere in the world? If not, then these cars must be charged manually -- which means only taxi companies with rented warehouse space and 60A chargers can run this sort of operation, not owners pressing a button on their app and letting their $60-100K car drive automatically all day.
In my opinion, I think some of your assumptions are correct, but the timeframe and the participation rate in the model are wildly optimistic.