#94 | Tesla Deep Dive | Full Self-Driving, Half the Story

Tesla in 2026 is in the middle of an identity crisis. Margins are compressing, the car business is shrinking, and Elon Musk is pivoting the whole story toward AI, robots, and full self-driving. We trace TSLA from the Roadster gamble to today's $1.6 trillion bet on optionality and what has to go right from here.

Transcript

SPEAKER_00 0:00

Welcome back to another TDI Premium Deep Dive by the Dutch investors. Today we are tackling a big one, a company that has been called a cult, a miracle, a fraud, and a car company, all in the same breath. Today's deep dive is about Tesla. But this deep dive isn't the Tesla of 2018 or even 2021. As we sit here in May 2026, Tesla is a company in the middle of a massive identity crisis. It is pivoting from being a hardware-centric automaker to what Elon Musk calls a physical AI company. We have to look at Tesla from all sides, because if we only look at one side of the coin with Tesla, we're not doing it justice. So let's talk about Elon Musk and Tesla. Let's dive in. Tesla was actually founded in 2003 in San Carlos, California, by two guys named Martin Eberhardt and Mark Tarpening. Eberhardt was an engineer and a sports car enthusiast. But he had a conscience. He was becoming increasingly aware of global warming and noticed that his favorite high-performance cars were, frankly, as guzzlers. He wondered if an electric car could be as fast as a Porsche 911. Now, at the time, electric cars were basically glorified golf carts. They were slow, ugly, and had the range of a cordless vacuum cleaner. But Eberhard took a test drive in something called the T0, an early electric prototype, and he realized that the potential was there. He and Tarpening formed Tesla Motors. But they ran into the classic hardware startup wall. They miscalculated the numbers, the financials. Hardware is difficult. It eats capital for breakfast and asks for more at lunch. Musk joined a year later as an investor, putting in $6.35 million. He wasn't just a passive check writer though. He became the chairman and eventually the architect of the brand. Elon Musk is someone who thinks differently about the world and looks in unique ways, in ways we don't really think. He has a wow mind. Musk pushed for a level of perfection and a master plan that was far more ambitious than the original founder's vision. Musk's strategy was a master stroke of storytelling. He published it in 2006. For the world to see, build a low-volume, expensive sports car, the roadster, to show Electric could be cool. Then use that money to build a mid-volume car at a lower price, the Model S. Use that money to build a high-volume, affordable car, the Model 3, and while doing all that, provide solar power. It sounds pretty simple now looking back, but in 2006, this was pure science fiction. Remember, this was before the 2008 financial crisis. It's a long time ago. Musk saw the mission to accelerate the world's transition to sustainable energy. This mission allowed Tesla to bypass the typical advertising spend of a Ford or General Motors. Why pay for TV ads when you have a CEO with a reality distortion field and a massive platform to tell the story for free? But we gotta ask, at what point does the story stop being a roadmap and start being a distraction? Are we there yet? Because by 2008, Tesla was literally days away from bankruptcy. Musk had to dump his entire personal fortune into the company to keep the lights on. I mean, talk about skin in the game, right? But it's also a reminder that Tesla has always operated on the razor's edge between miracle and collapse. So how does Tesla make money? As you may already know, Tesla earns the majority of its revenue from selling cars, electric cars. So, for most, Tesla is either a car company or a technology company. Both can be true at the same time. About 80% of its revenue comes from car sales. To say Tesla is just a car company wouldn't really do it justice. But if you generate about 70 billion in automotive sales, compared to just 12 and 13 billion dollars in energy and services, it is the majority of their revenue. Besides selling cars, they have a service segment. This includes everything else Tesla does, such as income from selling used cars, after-sale services not covered by warranty, paid supercharging, merchandise, and car insurance. Personally, we find the car insurance within this segment pretty interesting, because it can be a significant growth driver without Tesla needing to change the world. The energy segment includes all products Tesla offers for energy generation, such as the solar roof and energy storage, like the megapaks. Tesla is also working on robotaxis, human robots, AI, and several other futuristic projects. There's still a lot of uncertainty around these projects to make any assumptions, and therefore we kind of leave it out. Although there is a ton of optionality here. So let's discuss Tesla's mode. Elon Musk famously hated the word mode, competitive advantage. He once said that if your only defense is a mode, you won't last long, and that what matters is the pace of innovation. Now, Warren Buffett and Charlie Munger would probably have a few choice words about that. But we think modes are essential. But we agree that they aren't static. You have to widen them every year, throwing in crocodiles and pruanias to keep the competitors away. For a long time, Tesla's mode was its proprietary technology. They built their own battery packs, their own motors, their own chips, their own software. Tesla was busy vertically integrating. One of the biggest parts of the Tesla mode is something called the Gigapress. They have a massive manufacturing edge. Think of it like a giant die casting machine that makes the entire rear underbody of a car in one piece. Traditionally, you'd have to weld them together. 70 or 80 different parts. Tesla does this in one. It's not just cool engineering, although it is, but it is also a massive cost advantage. By reducing the number of parts, they reduce labor, they reduce the size of the factory, and they increase the speed of production. Economies of scale in action. Another significant advantage for Tesla is that supercharger network. Launched in 2012, this was Tesla's widest mode because it solved the range anxiety for people. Tesla has spent the last years opening its network to Ford, GM, and Rivian. Why are they giving away such a big advantage? Well, they are losing their switching cost advantage, you don't need a Tesla anymore to use the best chargers, but they are turning a cost center into a high margin services business. By early 2026, this segment alone reached $12.5 billion in annual revenue. There is a scene in Alice in Wonderland where the Red Queen says you have to run as fast as you can just to stay in the same place. Tesla's lead in hardware is being eaten, so they must innovate in software and services to survive. If they don't, they will just become another low-margin car maker. You could argue Tesla also has network effects. Tesla is a data-driven company. All decisions they make are based on data. The more Teslas are on the road, the more data Tesla collects. And it's a self-reinforcing network effect, giving Tesla access to an increasing amount of user data and driving data. And this helps Tesla innovate and provide an advantage in developing self-driving cars. Lastly, maybe Tesla even has a brand. You either love Tesla or you hate Tesla. It has a cult-like following. So the brand name is really there. So let's talk about compensation and the management. Charlie Munger famously said, Show me the incentive and I'll show you the outcome. Tesla's management compensation is the most extreme example of this in history of capitalism. Back in 2018, the board, which mind you, was full of Musk's friend and his brother Kimball, proved an extraordinary pay package. It was all or nothing. Musk would get options to buy 304 million shares, but only if Tesla's market cap increased from 50 billion to 680 billion, and they hit specific revenue and profit targets. At the time, Wall Street laughed. They said the milestones were impossible, but you guessed it, Musk hit all 12 of them by early 2023. The options became worth roughly 56 billion. But then a guy named Richard Tornetta, shareholder who only owned 9 shares, sued the company in Delaware. In January 2024, a judge voided the package. She found it ridiculous, found the approval process deeply flawed. She essentially said that the board was too close to Musk to negotiate at arm's length. Then it gets interesting. Tesla fought back. They asked shareholders to reincorporate the company in Texas and vote again on the same package. And in December 2025, the Supreme Court actually overruled the lower court and restored the package. So in April 2026, 304 million shares were being delivered to Musk, and the saga is officially over. Now this skin in the game and this thing with Elon Musk is a double-edged sword. On the one hand, you want your CEO to be an owner. Musk owns over $100 billion in Tesla stock. He isn't going to let the company fail because it would wipe him out. But on the other hand, the lack of an independent board means there is nobody to say no to his wild ideas. The same traits that make a founder a genius can also make them a liability. And let's be real, Musk's attention is split between Tesla, SpaceX, XAI, Neuralink, the Boring Company. He's running several different revolutions at once. Now the board has finally added a guardrail for a new shares to fest. Musk must remain CEO or a product development executive through at least 2028, but is four more years enough to solve the problem? I don't know. He now has another ridiculous package waiting for him if he manages to achieve it. Tesla Shareholders approved a new 10-year compensation plan for Elon Musk with an unprecedented equity package that could be worth up to, hold your breath, $1 trillion over the next decade, giving him another 424 million shares if all benchmarks are achieved. You might be wondering what does he need to do? To give you a rough idea, he has to increase Tesla's market cap to $8.5 trillion, reach 20 million in vehicle deliveries, secure 10 million paid self-driving subscriptions, deploy 1 million robotaxis, and deploy 1 million humanoid robots. He must remain CEO for at least 7.5 to 10 years, and he is required to present a formal succession plan for the company once he leaves. Well, now you know where the incentives are, you know what Musk is going to do. Now the heart of the investment thesis for Tesla is no longer the EV company alone. They describe themselves as a physical AI company. Now it's marketing fluff, but at the same time, they are serious. They are betting the entire farm on two things, the CyberCab and Optimus. In 2024, Musk unveiled the CyberCab, a two-seater with no steering wheels and no paddles. In 2026, Tesla announced that the first production had units had rolled off the line at the Gigafactory in Texas. Musk says that the goal is 2 million units per year. Now obviously, this is quite impossible to predict, but you know, it's Musk. There's also the regulatory wall. Musk thinks he can bypass the 2500 unit cap on non-standard vehicles by self-certifying that the cyber cap is safe. But vehicles without steering wheels are still illegal in most of the US and certainly in Europe. Waymo, however, owned by Google, is already years ahead with millions of unsupervised miles. To win in this AI war, you need compute power. Tesla is now building its own chips, but they are still dependent on the semiconductor web. During the 2020 chip shortage, the automotive industry was hit the hardest because they scaled back orders too early. Tesla was the only one that didn't hit the brakes, which is how they stole so much market share in 2021. But today the struggle is about EUV. To run the AI that powers robotaxis, Tesla needs the fastest chips on Earth, which are only made on ASML machines at TSMC factories. So you are only as strong as your weakest supplier, right? It's the reality of hardware. You can have the best AI in the world, but if you can't get a specific valve or high-end lens from a supplier like, I don't know, Zeiss, your factory stops. This is where things get really interesting. The humanoid robot. Musk has called this potentially the biggest product ever. The idea is that you have a robot that can do a human labor for $3 an hour. You can change the very structure of the economy if this works. In May 2026, Tesla is actually dismantling its Model S and X production lines in Fremont to make room for their first Optimus factory. They are targeting a capacity of 1 million units a year at Fremont and 10 million units a year at the new facility in Texas. A quick reality check. In early 2025, Musk predicted that they would build 10,000 robots by the end of 2025, and the actual number of robots? Just keep that in mind, right? The Optimus has over 10,000 unique parts. Manufacturing a humanoid robot at scale is a task of almost insane complexity. Boston Dynamics is also already shipping its electric Atlas robot to Hyundai factories, and Figure AI has a pilot at BMW. So Tesla is not alone in this race. Now let's briefly go over the financials. Tesla currently has a market cap of around 1.6 trillion, generating close to $100 billion in revenue. They employ over 134,000 people with a return on invested capital of just over 20%. The car sales are the majority of revenue, about 70 billion. It declined 10% year over year. Energy generation and storage actually grew by almost 30% last year to almost $13 billion. The same goes for services and other, growing about 20% year over year. It's funny how the core car business is shrinking, but the other segments are growing fast. Tesla Energy is definitely a hidden growth engine. But the real story is margin compression. In 2022, Tesla's operating margin was 16.8%. But in Q1 2026, it was down to 4.2%. They are caught in a brutal price war in China and Europe. Now these EVs and cars generate huge societal value, but no profit because of competition, same like the airline industry. So, is the EV industry becoming the new airline industry? They ended 2025 with $44 billion in cash. Their debt-to-equity ratio is basically non-existent. This gives them massive staying power. They can afford to lose money on car sales for years while they wait for the robotaxi miracle to happen. Tesla's net income fell about 50% in 2025, but their operating cash flow stayed flat at $15 billion, so they're still generating a lot of cash. But obviously, this is not loved by investors who pay a large multiple for Tesla. Now you know I like to look through the lens of evolutionary biology. Tesla sometimes reminds me of the following. There's this book called What I Learned About Investing from Darwin. In this book, they mention two types of errors. You have type 1 errors, you buy a bad business because you think it's good, self-harm, or a type 2 error, you reject the good business because you think it's bad, omission. Now obviously, if you didn't buy Tesla, you missed out on a massive return. But for every Tesla, there are hundreds of unproven business models that end up in the dustbin. Now, in this book, they use the Siberian Fox experiment as an example. Researchers in Russia tried to domesticate wild foxes by selecting for only one trait, tameness. And over generations, not only did the foxes became tame, but their physical appearances changed. They got floppy ears and curly tails. Now what does this have to do with Tesla? If you're selecting Tesla, you're basically selecting it for one trait, autonomy. By focusing everything they have, full self-driving and dojo, they are fundamentally changing their DNA. But just as those foxes became physically different, Tesla is becoming a different beast. It is no longer optimized for making high-quality, reliable cars. They just got rid of two, and the cyber truck recalls prove the quality isn't there either. It is optimized for AI training. The risk is that in the quest for autonomy, they lose the very thing that made them successful in the first place. The ability to manufacture cars people actually want to buy. If the tameness doesn't happen, you're left with a Fox, a car company, that can survive in the wild versus competition. If Tesla cannot get the full self-driving going, or the Optimus robots going, I mean, they are betting the farm. On those optionalities, we will never be right on the valuation of Tesla. Looking at it from a valuation basis like PE or price-to-free cash flow, it is way too overpriced. But that doesn't take into consideration all the optionality. When looking at a reverse DCF, way of seeing what the market is pricing in, assuming a 15% required rate of return and a terminal growth rate of about 4%, the market expects Tesla to grow free cash flow about 55% annually for a decade. If you believe that's possible, it's not so expensive. Looking at the incentives, he has to reach an $8.5 trillion market cap. If he reaches that, you'll still get a very nice 4 5 times your money if you invested today's market cap. So my final closing thoughts on the company. Because Tesla is almost the perfect company, as weird as It sounds. This perfection comes with very high expectations. And when expectations are high, then the chances of disappointment are great. In our opinion, Tesla has proven to be one of the best companies in the world. They deserve that spot. The vertical integration ensures that the company is operationally very sound. There's a clear focus on innovation. Innovation is embedded in the culture, the operational strategy, and even the compensation structure. Additionally, the management has repeatedly proven itself, and Tesla is led by a visionary. It is clear that Tesla is far from fully developed and it still has endless opportunities ahead. However, the fact remains that around 80% of its revenue comes from the sales of electric cars. It is also clear that this market will grow. But we must not forget that the automotive market is a very tough business. Competition is always lurking. China is always lurking. In our future predictions, we have tried to make some of our most realistic estimations for Tesla's future. Definitely take a look at our deep dive on tditerminal.com. But a lot needs to go right to achieve even a reasonable return. That's it for today. We hope you enjoyed this deep dive, and we'll see you in the next one!