Tesla’s Make it Or Break it Moment
Tesla is betting the farm - again..
Tesla is facing strong headwinds, in the last few quarters the company has been experiencing declining EV sales. This is primarily due to high interest rates impacting retail demand for EVs.
Recently Reuters tweeted that Tesla is scrapping their planned $25K affordable sedan a.k.a Model 2. Although Elon responded that “Reuters is lying (again)” he didn’t actually cite that anything was incorrect. On the contrary, a few hours after he announced that Tesla will unveil the Robotaxi - Tesla’s new model built from the ground up for being autonomous - on August 8. This can be seen as admitting that Tesla is scrapping their low-cost EV in favor of going all in on autonomous driving.
Thanks for reading Elon Weekly Digest! Subscribe for free to receive new posts and support my work.
Subscribed
Changing Growth Strategies
Tesla’s first growth spurt was achieved by successfully mass producing a mid-market vehicle in the shape of the model 3 and Y. This brought their market cap from $35B to at the peak $900B.
The playbook that Elon has been running in the previous years was to continually cut prices to increase demand for Tesla EVs, which in turn enabled to amortize the production fixed costs across more vehicles. This resulted in a cheaper cost per unit allowing for further price cutting, creating a flywheel. The introduction of the even cheaper model 2 would be this strategy on steroids. Investors were eagerly anticipating the production of the model 2 as a substantial growth engine for Tesla’s business.
This is all well and good, but Elon seems to realize that competing in a cheaper market segment would erode profit margins and eventually be a race to zero between the different EV companies. Elon stated in the past that his real competition are Chinese EVs and he can’t really compete with them on price since they can sell at aloss due to Chinese government funding them and their access to cheap labor. Elon is betting that the time is right to go all in into AI technology and using it to unlock FSD (Full Self Driving) and propell Tesla into its next growth phase. Achieving this will transition Tesla from being a car company which has low margins into an actual software company with high margins.
“Production Hell”
The previous growth Tesla experienced was enabled by the ramp up in EV production. The ramping process was painfully difficult and was nicknamed “Production Hell”. During this ramping process Tesla for several quarters was spending tons of capital into production lines that due to production complexity were initially producing very few vehicles. This period was associated with very negative PR, putting Tesla stock under pressure. There were also several high profile investors including Michael Burry who were short Tesla stock. There was a real belief that the company might not make it. Elon, who famously doesn’t shy away from a struggle, slept on the factory couch for several months and managed to eventually ramp up production. It is beneficial to remember the growth pains Tesla experienced during the previous company transition to better understand the difficulties that the company will soon be facing.
Tesla Transition Into an AI First Company
With the unveiling of the Robotaxi Elon is betting that getting to full Autonomy is viable. Looking into Apple’s “Titan” project - which was an attempt to create their own Robotaxi like vehicle - demonstrates how hard of a problem it is. Apple spent an estimated $10B on the project and had thousands of engineers working on it. After realizing how hard of a problem it is they pulled the plug and closed the project. If Apple, the world's biggest tech company wasn’t able to do it, it takes a lot of guts to believe Tesla will succeed.
To get an idea how capital intensive it will be to reach full autonomy, Elon stated in a recent tweet that Tesla will be spending $10B on training costs during 2024 alone:
Other difficulties Tesla will face before making their autonomous driving wide spread include regulatory hurdles, where each state will have their own laws and regulation on autonomous driving. I would also bet that every accident where an autonomous driving Tesla is involved will reach front page news. The early days of FSD vehicles will be associated with public fear, even though to get regulatory approval Tesla will need to prove that FSD is vastly safer than human driving.
The Prize
To get a sense of the opportunity it is worth looking at the TAM (Total Addressable Market) for FSD. It is estimated that 15.8 million workers work in the transportation industry, an approximate 10% of the total labor market. Not only will a driven mile be much cheaper for companies to use software rather than human drivers, there are secondary advantages to FSD including less accidents, and faster delivery times since FSD vehicles can work 24/7 with no need for breaks.
In addition to the labor market there are an estimated 239.24 million licensed drivers in the US alone. How much would an average adult pay to not need to drive?
Some of the ways that Tesla can generate profit are:
Build their own Robotaxi fleet that can undercut Uber and Lyft since there is no need for a driver.
Subscription service to their existing Tesla cars - which they already have today.
License out their FSD technology to other legacy car companies.
Legacy car companies rely on external vendors called “Tier Suppliers” to supply them with parts and software for their cars. Current Tesla competitors like MobelEye and Waymo require the use of LiDAR which uses lasers to see in poor visibility conditions. Tesla on the other hand requires only cameras, Elon famously said that “LiDAR is a fool's errand,” and that “Anyone relying on LiDAR is doomed” because of LiDAR’s tremendous cost.
Tesla FSD will have a moat since it will be extremely hard for other companies to reach Tesla’s level of autonomous driving because of their data advantage. Tesla has millions of cars they sold each with 8 cameras transmitting vast amounts of real world driving data. Every edge case that a Tesla car owner faces is another data point for Tesla’s AI model to learn from. This data acts as a moat since no other software company will have the immense amount of real world data to build a competing model.