TSLA: FSD and Q1 2024

TSLA is probably the most controversial stock in the market. I normally provide updates on my positions in my regular portfolio updates at the end of each earnings cycle. In the current cycle, which just began, that would be at the end of May. It’s a long time to wait given all that’s happened with Tesla recently. Therefore, I’m going to provide my thoughts on Tesla separately right now.

I first opened a position in TSLA position on August 2, 2023. I described my reasoning for entering into TSLA in my 2023-09-01 Portfolio Update: (2023-09-01 Portfolio Update - GauchoRico Stock Investing & Personal Finance):

“I bought a small position in TSLA (new position). The Company is outrageously expensive if all they ever do is earn revenue and profits from vehicles. My investment is based on the notion that TSLA will solve the autonomous driving challenge which would provide very high margin revenue from software sales. There’s also the possibility that some of TLSA’s other efforts (i.e. energy business or robot business) will pay off. I don’t expect that I would invest more than a 4% allocation in TSLA.”

For me the investment was a high risk, high reward type which could either lose value or do incredibly well. Therefore, I decided to keep the position small for if it loses value then I won’t lose a lot but if it does incredibly well then I don’t need a large position for the allocation to make a difference to my overall portfolio returns. During March 2024, when both the success and timing of achieving autonomy with FSD (full self driving) became more clear to me, I decided to sell all of my TSLA shares and purchase long-dated call options; this leveraged my position by increasing my upside by about 3x (without investing more capital) but put a time limit on achieving a positive outcome. March 2024 was when I began hearing more and more accounts from users about how quickly FSD had improved after TSLA opted to scrap their code-based algorithms controlling FSD in favor of a neural net approach that uses video collected by Tesla’s end users to directly train a model that controls the vehicle. Another factor in my decision to switch my position from shares to the more risky LEAPS was that the TSLA share price had fallen significantly during the first quarter of 2024. The stock was under pressure from a steady stream of negative news including Red Sea supply chain disruptions, an arson fire at their Berlin factory, higher than expected inflation and interest rates, several price cuts by TSLA, a 10% layoff, and negative YoY deliveries for Q1 2024. Going into the Q1 2024 earnings results, the expectation was that the results would be bad, and the stock price was at a multiyear low.

TSLA Reports Q1 2024 Results

Yes, the financial results were bad with respect to growth and cashflow. The Company reported a whopping negative $2.5B in free cash flow. Note that free cash flow had been very positive in each of the four preceding quarters. In Q1, Capital Expenditures ballooned to $2.77B. As pointed out on the earnings call, a large portion of the CapEx went to buying H100 GPUs from NVDA (love that because NVDA is the largest position in the portfolio!):

Elon revealed that TLSA doubled its AI training capacity in Q1 meaning that TSLA purchased about 20,000 H100 units between January and March 2024! Each H100 costs $40,000 so TLSA spent around $800M just on NVDA chips during the quarter. From this investment in AI compute, we can see that TSLA is putting its money where its mouth is in terms of going all-in on autonomy. Elon said that TSLA expects to purchase more GPUs and get to 85,000 (so 50,000 more H100s) by the end of 2024 (this is more good news for NVDA). I believe that TSLA has seen that their neural net approach is working and progressing so quickly so they are pushing hard to get to the finish line as fast as possible. As part of this effort, TSLA has also lowered the price of FSD from $199/month to $99/month and from $12,000 to $8000 for a one-time upfront payment (as an alternative to monthly payments); in addition, in April, TSLA started offering all its customers (who have version 3 or version 4 hardware in their Tesla vehicles) a free 30-day trial of FSD. Elon reported that there are 1.8 million cars that meet the criteria (version 3 hardware or better) to buy or try FSD and 900 thousand are currently using it (as of April 21, 2024). This is adding a tremendous amount of data for TSLA to train and improve its FSD model. The more data TSLA gets the more accurate and better FSD will get. Also, if TSLA gets more drivers using FSD then it can reach full autonomy more quickly. Here is a chart showing the cumulative miles driven with FSD:

That’s 300 million miles added in a very short time. Elon said that with the purchase of all those H100s from NVDA, TSLA is no longer compute constrained. Therefore, as more and more data comes in, TSLA can quickly update and improve the FSD models. In fact, Elon said that the Company has visibility into the progress and capability of the next two FSD versions, and he said the progress is amazing. For safety reasons, TSLA needs to spend some additional time testing these new versions before releasing them as updates to the entire fleet. Putting all of this information together, it sure seems like TSLA can see a clear path to autonomy, and the other information that I’m seeing points to a de-risked TSLA stock because the probability of FSD success is increased and the time to achieving autonomy is decreased. I’m not making any changes to my TSLA investment and plan to keep my leveraged (with call options) position in place.

Another quote from Elon during the earnings call which aligned perfectly with my initial reason for buying TSLA shares:

“So I mean, if somebody doesn’t believe Tesla is going to solve autonomy, I think they should not be an investor in the company.”
–Elon Musk on April 23, 2024 during the Q1 2024 earnings call

Well, I certainly agree with Elon on that one: without autonomy, TSLA shares are not even worth $100 per share. However, with autonomy TSLA is worth a lot more than the current price, and if TSLA can solve autonomy for cars, why can’t they also solve autonomy for humanoid robots, which, by the way, the Company is testing; TSLA expects to begin using their Optimus robot in their factories by the end of 2024.

GauchoRico

74 Likes

I often re-read posts from people here and I just kept coming back to this part of it. What I’ve gain from doing this is that I’m more clear that I focus strategically moreso that tactically, maybe that too general.

I think, because I’ve written here that I see Tesla as a ‘defensive move’ during a time when AI is expected to disrupt so much. And that I further explained that I believe Tesla’s Non-Auto segments are like a Call Option for me, with the Auto Portion as safe, I feel that a response to GouchoRico’s point is worth a post.

Simply put, It is clear to me that legacy auto has given up, they’re losing massive amounts of money on every EV and that the world is clearly moving to an 100% EV adoption. So, legacy Auto dead and Tesla Alive. Therein lies the safety.

Some may argue with the timing of legacies demise. My only counter is that I have posted here many times why we are currently in the vertical part of the hockystick regarding the rate of adoption and no one has provided any evidence to counter this. The talking heads on television are saying otherwise without any evidence. Yes, the automobile market is cyclical, both annually and when looking across decades; but, the trend toward EV adoption is only increasing throughout this overall landscape.

Or, maybe some if you are withholding information for some reason?

Please let me know, I’d really appreciate you sharing.

Best

Jason

12 Likes

This is correct as far as it goes but it ignores China’s push into EVs, that’s where Tesla’s competition in EVs is going to come from. But a much more important issue is AI competition. The recent events at Tesla point out an important pivot, from being a car maker to becoming a software and AI provider. The reason this is so important is because cars are a commodity business, a Decreasing Returns business, while technology is an Increasing Returns business. Apple is a great example, they don’t make the big money on iPhone hardware but on the iOS software. In Tesla’s case they will be selling FSD AI in cars and Humanoid Robot AI in the Optimist robot. Legacy automakers are essentially out of the race but we don’t know how advanced China is. The fact that China has allowed Tesla to use FSD is optimistic news for me because it shows that China wants to learn from Tesla.

Denny Schlesinger

Brian Arthur is one of the leading economists in the study of Complexity Economics that is replacing Classical Economic theory. His take on Increasing Returns:

12 Likes

I agree…somewhat. Wouldn’t it depend on how well Tesla protects its IP in the “learning” portion of these deals? How should we think about the second-order risk of China’s EV producers (or other companies) closing any technology gaps in the non-vehicle plays before Tesla fully monetizes them? Is there the possibility Tesla might plant seeds allowing others to catch up faster even as it tries to speed up its own release?

I’ll admit the tech is beyond my scope. I’m asking from more of a business perspective.

8 Likes

While I understand overall how the technology works I don’t know enough to answer the question but Tesla has opened it’s patents before so it might not even be an issue.

Denny Schlesinger

5 Likes

Fair enough. Thanks, @captainccs.

Either way, I don’t think this impacts the thesis TSLA is ultimately a non-EV play. It’s simply yet another factor in TSLA’s range of outcomes, including the still unresolved threat of Musk taking all the really cool toys to a different company if his pay package doesn’t keep him motivated enough to leave them with TSLA.

Watching this company is almost like a spectator sport.

12 Likes

Based on public information, the most likely way for Elon Musk to gain 25% control of Tesla would be through a combination of share buybacks by Tesla and the exercise of Musk’s existing option stakes. Musk currently owns about 13% of Tesla’s stock, and he has an option to buy an additional 8.6% stake. If Tesla were to buy back 5-10% of its shares over the next few years, it would reduce the total number of shares outstanding. This, combined with Musk exercising his options and paying the taxes, could potentially lead him to a 25% voting control of the company. This scenario assumes that the board approves a new compensation plan with performance-based criteria for new option grants.
I don’t believe Musk would ever give up on his Mission at Tesla, with or without 25% control. Nor do I want to see Tesla buy back shares.But, I do see Stocknovice’s point as being meaningful to the share price, at least during the next year or two.

9 Likes