Historic Opportunity Letter 2024

The Tesla FSD (Full Self Driving) Era Has Begun

Tesla’s FSD v12 was recently released. I had been using the previous version, FSD v11, daily for more almost a year now. FSD v11 was impressive. However, it did not wowed users to the floor because it required too many interventions from a human driver to either prevent an accident (referred as a critical intervention) or make the car less hesitant.

Last weekend I did more than 250 miles with FSD v12 without a single need to intervene in any way. I can assure you v12 represents an enormous leap forward in Tesla’s FSD capabilities. The best part is you do not need to take my word; you can find hundreds and hundreds of similar reviews about FSD v12 in social network X (previously known as Twitter).

The metrics from users are clear: miles per critical intervention for FSD v12 skyrocketed by more than 10x compared to the same metric for v11 (from 40 miles to more than 450 miles per critical intervention!). In addition, the car drives extremely smooth and does not hesitate. As some users stated, “you cannot tell whether it’s a human or a robot who is driving the car.”

The huge improvement from FSD v11 to v12 strongly suggests driving autonomy is closer than most people think. Especially when you combine it with the pace of iteration Tesla showed over the last month: an improved v12 version (v12.3.1, v12.3.2, etc) is coming out every two weeks. They still have to fix many corner cases (i.e. when an ambulance shows up) but the pure neural nets approach makes this work much easier. Previous FSD versions were coded; FSD v12 learns by watching videos of humans that drive properly. Every night, Tesla loads good driver behavior from their fleet of cars to their servers and make this happen. Today, no one else has the fleet of intelligent cars, the neural net models, and the computing to do this.

Robotaxi and Humanoid Businesses Among the Largest Ever

Tesla: % of Potential Business Value in 2030

Note: chart dimension is not meant to be precise. The point, however, is that Robotaxi and Humanoid business will likely dwarf Tesla’s current businesses. With a large software component, robotaxi and humanoird businesses are likely to have large profit margins.

Why is FSD so important? First of all, it will save tons of lives. There is about 100 deaths and more than 7000 injuries caused by car accidents every day in the US alone. Metrics show FSD v11 had one seventh the accident ratio human drivers have driving by themselves. I am sure FSD v12’s ratio is significantly better and will only improve from here.

Second, this technology opens the door to two of the largest businesses ever witnessed: robotaxi and humanoids. It’s important to mention that while robotaxi might be a few years away, Tesla already generated more than one billion of revenues from FSD subscriptions in 2023. This number is likely to multiply quickly from here. As important, the business will deserve a very large multiple on revenues given the size of the opportunity and its profit margins.

Tesla is currently offering a one-month FSD v12 free trial to every Tesla in the US and Canada. Regulation in Europe will allow FSD starting sometime between late 2024 and 2025. It is important to avoid thinking about robotaxi in black and white terms. The FSD business will grow meaningfully for Tesla before graduating into robotaxi status and become even larger.

As important, the probability for a legacy OEM licensing Tesla’s FSD technology has now increased dramatically. If we follow the example of last year’s opening of Tesla superchargers, we already know one thing: the minute one legacy OEM signs up, every single OEM signs up. To me this is a matter of “when, not if”. Every single Tesla comes with the ability to turn on FSD v12 or the latest version available. Once people become fully aware of the FSD v12 capabilities, who is going to want to buy a car without it? Remember, Tesla’s Model 3 and Y are already extremely affordable when compared to similar size cars.

Tesla's lead in the humanoid robot race, potentially the largest business ever, has also surged with the recent FSD advancements. Tesla’s FSD is basically software to operate a robot on wheels. There is an enormous overlap between the car and Tesla’s humanoid robot, “Optimus”. They both use a similar neural network training system, compute, and even many of the cameras and actuators.

The humanoid robot market is on the brink of an awareness and demand boom. A couple of weeks ago Nvidia hosted their annual event in which they showed about 15 different humanoid companies they are working with. Don’t get deceived by cool prototypes (same applies to electric cars). As Elon Musk says, “prototypes are easy, mass production is hard”. Eventually, the humanoid market potential could be in the billions of units and when it comes to mass production it is undeniable that Tesla is in a league of its own.

Short Term Noise Creates a Wonderful Setup

This great progress is happening at a very particular time in Tesla’s history. Today, investors are dumping the stock because deliveries of cars are down year over year. Mr. Market wants you to believe that the EV story and demand is in trouble. Please do not listen to this nonsense; nothing could be further from the truth.

Model Y and 3 sales (these two cars represent close to 90% of Tesla’s revenues) are down year over year for three specific reasons: 1) Tesla and the auto industry are annualizing a 20%+ drop in prices that started in Jan 23²; 2) in 1H24 we are also annualizing a large increase in interest rates that took place throughout 1H23; and 3) in 1H24 we are also annualizing the last important push from legacy OEMs into EVs³

Likely in a few months these short-term effects will be behind us and Tesla’s Model Y and 3 will grow again at a very healthy pace. Also, Tesla’s Cybertruck’s production is ramping extremely well and the next generation platform for cars under $25k is coming up in 2025.

These should propel Tesla units sold multiple times the current levels. Every month that goes by we are making great progress at leaving the headwinds mentioned above behind us while getting significantly closer to the Cybertruck’s and the next gen’s high growth.

At a 1.8% market share, we are very early in Tesla’s runway. High growth is pretty much  inevitable. Keep in mind that for every dollar spent on gasoline, only 20 cents of it is used to move an ICE (internal combustion engine) vehicle along the road. EVs, on the other hand, operate at about 87% - 91% energy efficiency. It’s a superior technology from almost every  angle you look at it: efficiency, safety, acceleration, agility, production cost (in Tesla’s case),  maintenance cost, FSD, etc. And the technology is only getting better. For instance, while today  it takes about 45 minutes to charge batteries to drive 300 miles using a Tesla supercharger level  3, in a few years this time will decrease to ten minutes or less⁴. 

It is not an exaggeration; the entire market will go to EVs over time. Moreover, Tesla’s  efficiency and track record shows that the company has every right to win a big chunk of this pie.

A Historic Opportunity

While I have no clue when will Tesla’s stock bottom, I am convinced of the following: in  a few years’ time we are all going to be asking ourselves why we did not buy more Tesla shares at current prices or even $220. 

KIG is not in Tesla for a small % return or even a large % return on any given year. We  are convinced Tesla is a transcendent company, is among the safest assets in the world (still looking for the other assets that are as safe as Tesla), and is likely to generate returns many, many times our invested capital. 

Tesla’s stock might be very volatile, but its probability of impairment of capital is  extremely low. Tesla has a pristine balance sheet, a competitive position that gets stronger by the  day, and, most important of all, a connection between their products and customers that few  companies in history ever managed to achieve. In addition, the world needs Tesla’s mission – to  accelerate the world's transition to sustainable energy – badly and Tesla has some of the best  talent in the world – their pick among the best engineers that think from first principles and are  nothing less than relentless – to make this happen. Call me crazy but I do believe it is imprudent not to invest enough in Tesla.

To make things better, the recent short-term situation with car deliveries is creating a  wonderful setup. As a smart individual put it, “today’s situation is analogous to the invisible  gorilla experiment⁵. Everyone is so focused on counting passes (sold cars, margins, etc.) that  they don’t realize a gorilla is passing by (incredible new products, FSD performance reaching  critical levels, etc.)”. 

As always, I thank you for your critical support. Please do not hesitate to contact us if  you have any questions.

Warm regards, 
Matias Sacerdote