Robotaxis: A Future Source of Income

How close are we to owning an autonomous car and renting it out as a Robotaxi?

The race to develop Robotaxis has seen huge investment, lots of competition and lots of concerns. However, big moves have taken place in the last few months with Apple pulling out of the race, and Tesla making big technological gains.

Elon Musk has claimed a Robotaxi could earn $30,000 / year and some people are already buying up Teslas in preparation to run their own autonomous taxi fleet. So where are we up to?

🦘 If you prefer to see how good the the technology is first, scroll to the bottom and watch the last video from Chansoo Byeon - it’s insane.

 💬 A quick background of how things are progressing

• A handful of US companies have been racing to develop autonomous vehicles: Tesla, Waymo, Zoox and Cruise

• Robotaxis have been deployed in a handful of U.S. cities (San Francisco, Phoenix, LA, Las Vegas) and China (Beijing, Shanghai) at a limited scale

• Until recently, the technology has not been advanced or cost-effective enough to scale - it has cost too much per car to challenge Uber

• Vehicles have required continuous remote monitoring/human oversight and have been much more expensive than human-driven ride-hailing services

• Experts were casting doubt on the potential of Robotaxis - with one University Law Professor claiming they can’t compete with an Uber driver who ‘makes less than minimum wage, has a midpriced car, and probably maintains it themselves’

• In 2023 the industry was hit with a major setback due to an accident involving the company Cruise (a human driver hit a pedestrian in San Francisco then a autonomous Cruise car ran over the victim and dragged them 20 feet before stopping)

• US auto safety regulators currently have investigations targeted at Waymo, Cruise and Zoox…. all for instances of irregular driving

🏎 Tesla’s Focus on AI has given it the Lead

• Tesla pivoted to base its technology on AI + cameras only (unlike other makers which use lidar sensors and require large amounts of coding)

• Tesla has 8 x cameras with visibility up to 250 meters installed around the vehicle

• Previously, custom written code (over 300 lines of C++) was needed instructing the car to slow down at speed bumps, stop at stop signs etc

• The FSD (full self-driving) software is trained on millions of hours of driving uploaded by Tesla drivers globally (with their permission) and is currently up to version 12

• Tesla’s customer-owned fleet drives more than 120 million miles per day, and over one million miles per day using FSD (US only)

• ARK Invest already estimates Tesla FSDs are 5x safer than human driven Teslas

• Its competitors are way behind - Cruise and Waymo have achieved 1 million miles each on public roads (compared to Tesla which achieves the same rate per day)

• All of this has means FSD software is training constantly on ever increasing datasets and improving exponentially (see videos below)

🚩 China is no slouch

• There are at least 10 Chinese automakers developing autonomous cars including BYD (China’s largest carmaker) and massive tech companies like Huawei and Baidu

• Xpeng plans to roll out an autonomous car for only USD$21,000 ($10,000 less than the price in China of Telsa’s Model 3)

• Musk visited Beijing in April 2024 to discuss the potential rollout of its FSD system and the possibility of securing government approvals for overseas transfers of data from Tesla cars (which is necessary for FSD to operate)

• In Q4 2023, BYD overtook Tesla as the world’s largest EV company with sales of 525,409 v 484,507 (although Tesla was still ahead over the full year selling 1.8m EVs v 1.57m BYD EVs)

• Companies like Baidu and Pony.ai have already launched limited robotaxi services in Beijing and Shanghai

🍎 Apple Pulls out of Race

• In February 2024, Apple officially cancelled its autonomous car project

• Apple first started talking about developing a car in 2014 and is estimated to have spent ~USD$10 billion on the project

• The project was plagued by delays and downgrades - initially targeting level 5 (full self-driving), then level 4 and eventually level 2

• The cancellation timing also coincided with a plateuing in global EV sales, and increasing EV competition

• While Apple was spending big, their miles tested were still well behind competitors

• Some have speculated that as Tesla’s FSD miles have exploded, Apple realised it was unable to catch up

👓 What about UBER?

• Uber originally had plans to develop its own autonomous cars, but has now taken a partnership approach

• Roughly 2 years after selling its own autonomous division in 2020, UBER signed a 10 year deal with Motional Inc (owned by Hyndai) to provide driverless rides to its users

• Uber has also formed a partnership with Waymo (owned by Google) to offer driverless rides in Pheonix, Arizona, with the service launched on 23 October 2023

💳 The Road to Monetisation

• In 2016 Musk first talked about being able to one day buy a Tesla and then add it to the fleet so it could drive when you don’t use it

• At a Tesla investor day in 2019, he then mentioned ‘robotaxis could generate owners up to $30,000 a year in profit’

• Musk has announced the Robotaxi business will launch on 8 August 2024

• Cern Basher discusses this topic in depth in this youtube interview with detailed revenue and profit forecasts (see charts below)

• He estimates that if Telsa takes 30% of revenue, then an individual owner could earn between USD$37,676 - $80,160 depending on how much it is used (ie. 30% up to 65% each day)

• Once costs are taken into account he suggests profit could be between USD$20,604 - $53,025 (which could mean being able to pay back the cost of a brand new Tesla in 1 year)

• ARK Investments puts forward a bull case for the impact on Tesla’s share price putting it at $2,500 (it is under USD$200 now)

• In addition to driving revenue, it could also be possible to earn advertising revenue, subscription revenue, deliveries or other services like commission paid by preferred locations (ie. if the robotaxi recommendeds one restaurant over another)

• One company in Europe, Mister Green, is committed to becoming the first autonomous fleet business using Tesla’s Robotaxis

• The global expansion of Robotaxis will depend on regulation - no countries have national level regulations, but certain cities have allowed them in pre-defined areas. As the technology is recognised as safe, this should expand.

👀 Other thoughts

• If widespread adoption is reached you would think overall car ownership will plunge due to availability of rides (except for those owning Robotaxis), the importance of garages / carparks will diminish and the utilisation of short range public transport may also drop

• Longer term, the robotaxi network will get more and more efficient with less manually driven cars on the road (fewer humans getting in the way, making mistakes etc)

• Imagine when all cars are available on an ‘as-a’service’ basis and you can pick and choose the model depending on your need / mood. Simply order a truck, a convertible, or a sports car. Or one step further into the future - order car where you can work, where you can sit 4 people in a lounge setting, where you can play computer games etc… (look at the Zoox design)

• You can see elon Musk himself launching FSD in August 2023 and going for a drive:

• Compare the above clip to a user demonstrating his drive recently (May 2024) and see how much it has improved:

This truly is one of the most high potential ways we will earn money in future, and it follows up my last article on the possibility of owning robot workers. It is all part of the same trend.

I hope this was interesting and thanks for reading! Please get in touch if you have any qustions or feedback.

Jack