Tesla's Q3 Summary: Slow progress in CT, slow construction Mexico , Musk has bankruptcy "phobia".

1. Key points from executive statements:

1. Accumulated autonomous driving mileage exceeds 520 million miles; deployment of 10,000 H100 GPU clusters has been completed.

2. FSD V12 is end-to-end AI, and significant investment has been made in autonomous driving. Success in this field could make Tesla the most valuable company in the world. There is almost no limit to the scale of autonomous driving vehicles and robots in this business.

3. Energy is now the highest-margin business, contributing $500 million in the third quarter.

4. Cybertruck (CT) production capacity is challenging, and cash flow turnaround is also difficult. However, for innovative products like this, it is normal.

5. Two emphases: a) Goal: to ensure positive cash flow while ramping up production capacity, with a sales target of 1.8 million vehicles; b) Tesla's AI is the best in the real world and has the most powerful AI team globally.

6. R&D expenses are increasing: a) CT prototype development and trial production costs; b) investment in autonomous driving, Optimus, and DOJO. Continued investment in these areas will lead to short-term growth in R&D costs, but positive operating cash flow will be ensured. As of now, operating cash flow is $8.9 billion and free cash flow is $2.3 billion.

7. Buying a car usually requires borrowing money, and with the significant increase in US interest rates, adjustments need to be made to the actual monthly payment amount. We have introduced a new car purchase plan in the US, selling cars through leasing. The standard monthly lease for Model Y is $399.

Details are as follows:

2. Q&A Analyst Questions and Answers

Q: How many Cybertrucks do you expect to deliver?

A: CT is a completely innovative product, not only in terms of product innovation but also in manufacturing methods. The higher the level of innovation, the more difficult it is to predict the results. It is difficult to determine the ramp-up of CT production capacity. The overall goal is to achieve an annual production of 250,000 vehicles, but it will be challenging to reach that in 2024. There may be hope at some point in 2025.

Q: Can you provide an update on the 4680 battery?

A: The production of 4680 batteries in Texas has increased by 40% quarter-on-quarter, with a cumulative output of 2,000 cells. Texas is currently the main 4680 factory, and both production capacity and efficiency have improved. Waste in the last quarter decreased by 40%.

The energy density of CT batteries is 10% higher than that of Model Y cells. In order to simplify production and focus on output, both production lines will be converted to CT battery production lines this quarter. In the next three quarters, all three production lines in the first phase will continue to ramp up production.

The construction of four additional production lines in the second phase is underway. These new production lines will further improve efficiency, with the goal of starting production capacity in the second phase by the end of 2024. The Kato Road line in California is being reconfigured to prepare for the design of the next-generation battery, which will be a step ahead of mass-produced batteries, enabling faster iteration and smoother implementation of new designs.Q: What is the production situation in Berlin and Austin?

A: For the factories in Berlin and Austin, the current focus is actually on maximizing the output of the existing production lines through a focus on efficiency improvements. High quality and reducing unit costs will be just as critical as increasing production volume.

As for Mexico, we are simultaneously working on infrastructure and factory design, as well as engineering and development for new production. However, before fully constructing the Mexican factory, we need to confirm the global economic situation. Currently, there is significant pressure from high interest rates, which poses a significant burden on monthly car payments.

Even if we start advertising to boost sales, it can certainly help to some extent. However, if the monthly payment for purchasing a car is too high, ordinary people still won't be able to afford it and won't buy it.

Q: Can the sales target of 2.3 million in 2024 be achieved without the release of new Volkswagen models? When will we return to a 50% annual compound growth rate?

A: There are many uncertainties in 2024, and the core focus is on maximizing the efficiency of selling cars. No one can achieve a permanent 50% compound growth rate, but what can be certain is that we will grow rapidly and run faster than any other company in the world.

Q: What is the progress of Robotaxi?

Robotaxi is definitely about autonomous driving. Autonomous driving is exciting, mainly because of its end-to-end neural network self-driving software, which allows me to wander around Austin without taking control. It will also allow Optimus to do useful things and ultimately liberate human labor. In this regard, no one can do better than Tesla.

Q: Why is the price of FSD decreasing?

A: The price reduction is to allow more people to try it. The price of FSD will increase proportionally with its value, and the current price can be considered temporary.

Q: Regarding Level 3 autonomous driving, Mercedes has expressed that they will take legal responsibility first. What about Tesla?

Many people believe that we have legal responsibility based on lawsuits. Whether we like it or not, we will definitely not escape this predicament.

But the Mercedes system is limited to a few roads in Nevada and certain cities in California, and it cannot work in snow or fog. The speed is only 40 miles per hour when there is a car in front. Our system is comprehensive and can operate in any scenario, so our system is definitely better.

Q: Will Optimus start production at the Gigafactory next year?

There are no details now, but there will be regular updates. However, based on our progress, a year ago it might not have been able to walk, but now it can do yoga, so maybe in a few years we will be able to do ballet.

Q: What is the progress of FSD's international deployment?

The more places it is deployed, the more difficult it becomes. Making it available everywhere will take longer in each place. Therefore, it is currently only deployed in North America. In addition, international deployment requires regulatory approval. But in North America, we can take on a certain amount of risk to deploy it. At least you are responsible for your own risks. So, we will only go through the cumbersome approval process when we are very certain.

The learning speed of AI is like a logarithmic curve, with rapid initial improvements. As long as it can be inferred, it will continue to improve itself and drive at a fairly fast pace.Because the improvement rate curve is logarithmic, the asymptotic line will not occur multiple times. We can view the capabilities of AI as a series of stacked logarithmic curves, with each curve higher than the previous one. By continuously stacking logarithms, they will eventually be left behind.

Q: It sounds like it will take until the third year for CT to truly ramp up production. Is the next generation vehicle platform a similar situation?

A: It's not the third year of production, but 18 months. Although it's three years, it's actually only 18 months, and the effort put in during this period is unimaginable. The network has unique complexity. It takes a long time for new things to reach a market size that allows them to thrive, which is the nature of new things.

However, CT is a "self-digging grave" type of innovation, but making a high-volume, low-cost small car actually requires a lot of tradition. It doesn't require a great invention. We are just trying to simplify the vehicle as much as possible, making it more streamlined, automated, and cost-effective.

Q: Before Robotaxi is introduced, will a simplified version of FSD be sold? Has its financial contribution been calculated? How will it be priced to increase usage?

A: The economic benefits of a fully autonomous vehicle are astonishing. Existing vehicles are only used for about 10-12 hours per week, which means that out of the 168 hours in a week, only 10 hours are actually used for driving if you drive an average of one and a half hours per day. You also have to pay for parking fees, insurance fees, maintenance costs, and other expenses.

If a car can drive fully autonomously, its utility can increase many times over. For example, if it can be used for 50 hours a week, but its cost is similar to before, the efficiency has greatly improved. We are a hardware company, with software gross margin.

Q: How does the gross margin change? How much does factory idle time affect it? What benefits will the increase in production in Berlin and Austin bring? The gross margin for energy storage is impressive, how does it affect the gross margin?

A: Factory idle time has some impact on the gross margin for Q3; but our other factories are operating well and contribute significantly to the profit. In fact, one of the new factories has a very low unit cost, close to the level of our California factory.

In terms of energy profit margin, the deployment of Megapack is a key driving factor there, and it has performed well, which is positive for this quarter. However, the deployment of Megapack is a bit unstable. So, when we try to push this product into other markets, there will be pressure.

Q: The company has reduced the cost of each vehicle by an average of $2,000 over the past year. How much does this specifically affect the product profit margin? And what is the specific timeline for the next generation of products?

A: There are many factors that affect the product profit margin, and currently we can only try to control costs to see the subsequent effects. The company completed factory updates in the late third quarter, but the fourth quarter is also the stage where the updated factory has just started running, and the factory has not yet reached its ideal state, so there is short-term uncertainty. In addition, the CT ramp-up will also have a negative impact.

At the same time, due to the uncertainty of the macroeconomic environment, it is currently difficult to provide specific forecast results, but the company will continue to make adjustments. The launch date of the next generation of products cannot be disclosed.Q: How do you view the price elasticity of products in the current macro environment?

A: I believe there is a significant price elasticity. To be honest, if the cost of a RAV4 is the same as ours, no one would buy a RAV4. I want to specifically mention the tax deductions. For ordinary people, many of them do not have bank deposits of 7,500 or 10,000.

Most people earn money and spend it immediately, even carrying debt. For someone who earns 200,000 a year, it may be difficult for them to understand what it's like to have an annual income of 50,000 to 70,000. Therefore, for many people, even if they can recover the cost through tax deductions after 18 or 6 months, their taxes may not be enough to offset the cost. So, for these people, what they care about is how much money they need to pay immediately and how much they need to pay monthly. That's the key issue. From this perspective, our vehicles are still much more expensive than a traditional fuel vehicle.

For many people, price reduction is a necessity, not an option. Look at how much we have reduced the price on the Model Y, and then consider how many people have paid more interest due to rising interest rates. The price of the Model Y has hardly changed. Moreover, after interest rates rise, even if some people still have jobs and can afford to pay, banks are more cautious about lending and are unwilling to lend.

Many regional banks are already struggling, even century-old stores like CS have closed down. In addition, there is great pressure on commercial real estate, and it is estimated that more banks will close down. Credit card loan interest rates are already over 20%.

Q: There are reports that the company will offer radar as an option in some models in China, is this true?

A: Radar is not included. We have included radar as an experiment in the design of Tesla in Model S and Model X, that's all. But we have no plans to integrate radar into Model 3 and Model Y because humans can drive very well. In fact, an excellent human driver can drive with astonishing safety based on vision alone. Cars can observe all directions simultaneously through vision, which is beyond the average safety level of humans because we don't have eyes in the back of our heads. And computers never get tired or distracted, nor do they get drunk.

As for radar, what really matters is how much it affects the probability of accidents. In order for radar to be effective, it must be able to perform pure radar braking or pure radar operations, otherwise there will be ambiguity between vision and radar. Therefore, Tesla has turned off radar in vehicles because radar generates more noise than signals.

Now, the (4D millimeter-wave) radar designed by Tesla is a high-resolution radar, but its actual effectiveness still needs to be observed.

Q: What are the factors that will lead to a decrease in vehicle costs in the coming quarters? Is it an increase in factory utilization or a decrease in material costs? Can you provide data to prove that costs will decrease over time?

A: In terms of costs, we are always committed to reducing the manufacturing costs of our products and improving production efficiency. This involves engineering innovations and working together with supply chain partners to automate some production lines, eliminate bottlenecks and high costs, and ensure that factories can obtain components in a timely manner.We are not only focusing on one aspect, but controlling costs comprehensively. We must control costs in every detail because the cost of each component will affect the final product price.

Take a $40,000 car as an example. If there are about 10,000 components, the average cost per component is $4. To reduce costs by 10%, an average of 40 cents per component needs to be saved. This is an ongoing process. We are pursuing a set of measures internally called "cost battle". We will examine each line to see how to make it better. It is a laborious and persistent battle, just like digging a tunnel with a spoon.

We have also made some factory upgrades, and it is expected that production will increase, which will also bring higher production cost savings. However, on the other hand, we will launch new products, such as the Cybertruck we mentioned before, which will bring new challenges and opportunities. These are the practical problems we are working hard to solve.

We need to find a balance in different aspects to ensure that our production runs efficiently and achieve cost control and savings. But we will also improve the quality and functionality of the car very strictly. Only fools would reduce the cost of a car by lowering its quality or removing features.

Q: Will the plan for the Mexican factory be indefinitely postponed until the economy improves?

A: No, building a factory in Mexico is certain. It took a long time to choose, and the location is also good. It's just a matter of timing, and interest rates must drop, otherwise the purchasing power issue cannot be solved. But I can't see this issue clearly. If you tell me what the future interest rates will be, I can tell you when we will start. However, we will still start preliminary construction next year.

I hope to make fewer mistakes. If I seem stubborn, I apologize deeply because I have suffered from post-traumatic stress disorder (PTSD) since 2009, which was a very difficult experience.

Order registration also has some cyclicality. If there is economic uncertainty, people will hesitate to buy cars, so product companies perform very well in periods of economic prosperity and poorly in severe economic challenges. Especially when there is an ongoing war, it affects your emotions. People are watching the news and seeing wars happening in different places. Buying a car in the city is often not the top priority.

Risk Disclosure and Statement of this article: Dolphin Research Disclaimer and General Disclosure