I would taken a slightly different approach.

Let’s assume the market for driveless cars consists of 2 car groups in year 2020:

– Group 1 – low price driveless cars for carsharing companies .

These cars will have minimum functionalities and their only job is to move people around. Tesla Model 3 costs $35,000 in 2017. This is a car that has self-driving capabilities and some extra features that is not needed for an Uber / Lyft car. E.g. LCD monitor. Let’s assume an Uber / Lyft driver-less car costs $30,000.

Number of Uber / Lyft cars to be needed is estimated as followed:

Let’s say that in the U.S. 10% of Americans have signed up to Uber / Lyft. And they take 5 rides per week with it. Each ride takes 10 minutes. And currently, an average Uber / Lyft car works for 40 hours per week with 75% utilization. This means there are a total of

10% of 330million = 33million car sharing users

33 million users x 5 rides per week x 10 minutes per ride = 1650 million minutes of rides per week

Let’s assume Uber / Lyft will complete 10% of the rides with self-driving cars in year 2020. So, 10% of 1,650 million minutes = 165K million minutes per week of ride sharing will be completed by driver-less cars.

Let’s say that an average self-driving car works for 22 hours per day at 70% utilization. In other words, one self-driving car can complete 22 hours x 70% x 7 days x 60 minutes = 6,600 minutes of ride per week.

Number of cars needed to serve: 165K million minutes / 6,600 minutes = around 250K cars

That means Uber / Lyft will buy a total of 250,000 driverless cars. Total market size for ride-sharing driverless cars will be

250K x $30,000 = $7.5 billion

– Group 2- Personally owned cars.

Let’s say Tesla and Google will launch their self-driving cars by 2020. Tesla currently sells about 50,000 cars per year and is expecting to significantly ramp up production over the next two years. Let’s say Tesla will sell 500,000 of cars in year 2020.

Tesla cars are currently priced at $35K, $70K, $100K, and $120K. Given most of the volume is expected to come from the Model 3, I will assume the average price of a Tesla car to be $40,000 by year 2020. In other words, I expect the total market of Tesla cars to be

500K x $40K = $20 billion

Google has not started making any cars yet so chances are that the number of cars produced by Google will be small by year 2020. Let’s assume it will be 10% of Tesla, meaning the total sales by Google will be $20B x 10% = $2B

Now, let’s calculate total market size of self-driving cars in the U.S. : $7.5B + $20B + $2B = around $30B

Key assumption made: carsharing companies are making their own cars and are not buying from Google and Tesla. If they were buying from Google and Tesla, then we just needed to estimate the volume of cars sold by Google and Tesla. I have also not included the market size of used Tesla cars that will be upgraded to support self-driving.