The idea of a driverless fleet of electric cars powering an autonomous ride-hailing service is a popular one right now – especially since Tesla announced that all its new vehicles are now equipped with the necessary hardware to achieve full autonomy through software updates.

Tesla’s upcoming service even has a name now: ‘Tesla Network‘. I think it could be a good opportunity to revisit a study from earlier this year that tried to estimates the cost of what we know now as a “Tesla Network ride” in a Model 3.

As it turns out, the study, which was published back in April by T. Donna Chen of the University of Virginia, Kara Kockelman and Josiah P. Hanna of the University of Texas at Austin, is still holding up so far based on the price Tesla is now selling its “Full Self-Driving Capability’ option ($8,000).

The study looked at the cost structure of a fleet of what they called ‘Long Range Shared Autonomous Electric Vehicles’ (LRSAEV), and they decided to use a Tesla Model 3 as an example since it was expected to be released with self-driving capability at the time. And now we know that it will at least have the necessary hardware to achieve full autonomy through software updates.

They looked at a particular market, 100 mile by 100 mile grid over a simulated Austin, Texas, and they tried to figure out what it would take for a LRSAEV fleet to take a 10% market share of all the trips in this grid. They estimated that you would need a fleet of 31,859 Tesla Model 3s and they even accounted for chargers. The study suggests that you would need 1,517 Level III chargers to support the fleet.

Here’s the structure of the rides in the market and different possible cost structures of the fleet:

Based on the mid-cost structure, the researchers arrived to a cost of $0.479 per mile plus an additional $0.184 per mile for general administration costs for a total costs of $0.663 per mile. That is significantly less than any other ride-hailing or car-sharing service in the same market and it is even competitive with private car ownership.

Here researchers elaborated on the results and compared them to other available services in Austin:

These costs are on the low end of current manually-driven free-float carsharing services such as Car2Go, which charges roughly $0.70 to $1.23 per mile in Austin, Texas (assuming trips are between 2 to 10 miles and travel speeds are between 15 to 35 mph). Under this pricing assumption, SAEV [Shared Autonomous Electric Vehicle] users would pay roughly 21 to 49% of what is currently charged by transportation network companies like Uber and Lyft (whose equivalent per-mile pricing is $1.50 to $3.18 in Austin). In fact, these costs are competitive with AAA (2014) estimates of average costs of private vehicle ownership, which ranges from $0.40 to $0.95 cents per mile depending on annual mileage and vehicle type, suggesting that availability of a SAEV fleet can have significant impacts on private vehicle use (and ownership), particularly for low-mileage households.

Now we know more about the Model 3 and Tesla’s proposed car-sharing fleet of self-driving vehicles since the study was published. The cost structure could actually be closer to the ‘low cost’ version since we now know that Tesla is currently selling the necessary system to achieve full autonomy at $8,000.

That would price the Model 3 without any other option at $43,000 while the low-cost option is at $45,000 and mid-cost at $50,000.

Additionally, we now know that most of the vehicles in Tesla’s shared fleet of self-driving cars, Tesla Networks, will be owned by customers, which could impact the cost structure too.

Nonetheless, it’s an interesting starting point and it shows a clear path to significantly lowering the cost of personal travel using a shared fleet of long-range autonomous electric vehicles.