- Tesla is leading in developing Full Self-Driving (FSD) via vision-based neural network, and is the front-runner to bring FSD to the market.
- Tesla started to use its own chips in Models S, X, and 3 since 2019, and such hardware can be upgraded to FSD simply via software update.
- Tesla can leverage its real-world data, chip design, and financial capability to strengthen its competitive advantages in the FSD technology.
- Tesla has an enthusiastic fan base and strong branding power, which can further help build a strong economic moat.
- Looking forward, Tesla’s technological advantage and branding power will help it build a strong economic moat. Thus, I rate Tesla a buy.
a business with a wide and long-lasting moat around it… it could be because of its position in the consumer’s mind, because of technological advantage.
– Warren Buffett at 1995 Berkshire Hathaway annual shareholder meeting
A trendy business school professor, who I know is an Apple fan, sent out an email to a group of colleagues that he just ordered a Tesla Cybertruck online. He has an Apple Watch, a few MacBooks, an iPhone, and an iMac in the office. There are no formal statistics regarding the overlap between Apple and Tesla fans. In a lot of aspects, nevertheless, the similarities and parallels between Tesla and Apple are easy to spot. Both are in the high-tech industry; both are headquartered in California; and, both have enthusiastic fan bases.
Different from Apple earning a consensus buy rating though, Tesla is a love-it or hate-it stock. When asked the question “Would you invest in Tesla?” in a recent interview, Warren Buffett simply answered “No.” With Tesla’s soaring share price recently, you might wonder “Should I invest in Tesla?” Contradict to Buffett, I rate Tesla a buy and explain my reasoning in this article. The premise of my thesis is that Tesla meets Buffett’s two criteria for an economic moat, namely technological advantage and branding power. Specifically, in this article I will first provide an overview of Full Self-Driving (FSD). Next, Tesla’s unique approaches to develop FSD and their implications will be discussed. Finally, I will detail why technological advantage and branding power can help Tesla build a strong economic moat.
Table of Contents
FSD Overview
To understand how FSD can help Tesla build an economic moat, we need to have a broad-scope understanding of the technology. For this purpose, I provide an incomplete list of FSD developers in Northern America below.
| Developers | Operating Car Numbers | Own or Partner | Manufacturing Stage | FSD Progress | Chips | Sensors |
| Tesla | 989,861(825,970 with autopilot 2/3) | own | own factory, mass-production | only Autopilot | own chip | camera without lidar |
| Waymo | about 600 in total (as of 2019) | own | order cars from Chrysler and Jaguar | FSD, but in a controlled geofenced environment | Intel | camera with lidar |
| GM (acquired Cruise) | about 180 (as of 2017) | own | own factory | tested in a few cities | not clear | camera with lidar |
| Ford | – | Argo AI | – | – | not clear | camera with lidar |
| Fiat Chrysler | – | Aurora | – | – | Nvidia | camera with lidar |
| Rivian | not clear | own | own factory (estimate to be on the market in 2020 or 2021) | – | not clear | camera with lidar |
| Zoox | 50 test vehicles | own | not clear | – | not clear | camera with lidar |
| Uber | not clear | own | order cars from Volvo | tested in a few cities | Nvidia | camera with lidar |
(Table 1: Overview of FSD Developers; Credit: The author)
(Figure 2: Tesla Deliveries and Autopilot Hardware; Source: Lex Fridman)
I provide summaries below:
- Waymo has achieved FSD, whereas Tesla has not. At first glance, Waymo is leading in the FSD race. However, the caveat is that Waymo has been operating in a controlled geofenced area.
- Tesla has delivered approximately one million cars (see Figure 2). More than half of the delivered Tesla cars have autopilot 2 or 3, which can provide data feedback to help Tesla develop FSD. In contrast, Waymo has much fewer cars on the road.
- Besides Waymo and Tesla, there are two groups of players in the field. The three major car manufacturers entered the FSD space via either partnerships or acquisitions. GM has acquired Cruise; Ford has a partnership with Argo AI; and, Fiat Chrysler partners with Aurora. Another group of players are Rivian, Zoox, and Uber.
- For sensors, except for Tesla, all developers use lidar. To be clear, most developers also use cameras, radars, and ultrasonics (see Table 2).
(Figure 3: Tesla Dynamics and Economic Moat; Credit: The author)
Analysis: Tesla’s Unique Approaches and Their Implications
Prior to the analysis of economic moat, in this section I want to discuss Tesla’s two unique approaches to develop FSD. Such a discussion can provide background knowledge for the analysis followed. Further, I have noticed discussions on Seeking Alpha about the uniqueness of Tesla as a tech company. For instance, one user commented “To say TSLA is a tech company is not the whole story as there are plenty of tech companies with more understandable stock prices.” I believe this section can provide insights into such discussion as well.
Analysis 1: A progressive approach providing sustainable R&D support
Different from Waymo, Tesla does not seek to achieve FSD instantly. Waymo is the first and only one so far applying FSD into actual operating service. Waymo started to test its vehicles in Phoenix since 2017. In early 2018, Waymo launched a limited public ride service. To achieve FSD, Waymo puts a lot of constraints on such service including limited customers, limited periods, and limited areas. Such a constrained service can only generate limited revenues. Obviously, Waymo doing so is not for revenues, but rather to gain experience and data. On the flip side, it means that Waymo has to keep getting investment from either Google or outside investors. Indeed, they did. In March 2020, Waymo announced that it raised its first external funding of $2.25 billion.
(Figure 4: Tesla’s R&D Spending; Source: Statista.com)
The advantage that Tesla does not pursue such FSD is to have sustainable financial support for its R&D. Since such constrained FSD is not suitable to sell cars to consumers, Tesla instead sells its electric cars with autopilot – a driving assistant program (I will discuss more the transition from autopilot to FSD later). In the recent quarters, Tesla has been making profits. As shown in Figure 4, Tesla spent more than $1.3 billion annually on R&D from 2017 to 2019. Thus, Tesla has outspent Waymo in R&D. Sustainable and strong financial support can allow Tesla to attract the most talented scientists and engineers. According to Universum, Tesla and SpaceX are ranked the top two ideal employers among engineering students.
Analysis 2: Lidar or not and its implication
To “see” the environment, cars must have sensors, which can be lidar, radar, camera, and ultrasonic. Most developers use a combination of them (see Table 2). Developers including Tesla typically use cameras to read traffic lights and road signs. Beyond that, cameras can be used to measure distance and depth. Tesla thus believes that they can just focus on cameras to develop FSD. Except for Tesla, other developers do not think that cameras will be enough, and FSD needs help from lidar. Lidar sends thousands of laser pulses per second to measure the distance between surrounding objects and the vehicle to create a detailed 3D map. The most significant disadvantage of lidar is its price, which is in the range of $5,000 to $10,000 per unit. Thus, if Tesla can achieve FSD without lidar, all other developers need to play catch up to follow such a technical route, due to the obvious cost advantage.
| Sensors | Brief Explanation | Advantages | Disadvantages | Who Using |
| Camera | use image classification + object localization (similar to smartphones that can automatically detect human faces when taking photos) | (1) relatively cheap compared to radar and lidar; (2) can read traffic lights and road signs | (1) can be impacted by weather and lighting conditions; (2) require a lot of processing for extracting information | most developers |
| Lidar | send out laser beams to measure the distance between surrounding objects and the lidar | (1) can form a detailed digital 3D environment;(2) can reach a distance range of 100 to 250 meters | (1) high cost: $5,000 to $10,000 per unit; (2) do not do very well in the snow, fog, rain, and dusty weather conditions | most developers except for Tesla |
| Radar | uses radio waves to detect the environment | reliable under fog, rain, and snow; low cost | less accurate than lidar | most developers |
| Ultrasonic | send out short ultrasonic impulses that are reflected by obstacles. Such echo signals are received and processed. | can work normally in bad weather (rain, snow, dust) | can only reach a short range of 6 meters | most developers |
(Table 2: FSD Sensors; Credit: The author)
(Figure 5: FSD Sensors; Source: Engineering.com; Edit: The author)
(Figure 6: Argo, Uber, Waymo, and Cruise using lidars (i.e., ones in the red circle); Source: Internet; Edit: The author)
(Figure 7: Tesla’s sensors, including cameras, radars, and ultrasonics; Source: Tesla website)
Will Tesla be the first one to bring FSD to the market? I believe Tesla is well-positioned to achieve it. Tesla is confident about its no-lidar approach, evidenced by it selling FSD software to Tesla car owners at $7000. In the 2020 Q1 conference call, CEO Elon Musk even said Tesla could offer a subscription-based FSD service by the end of 2020. Further, since none of the developers has achieved FSD yet, real-world data is needed to provide feedback, regardless of the technical route. In this aspect, Tesla stands out. With many Tesla cars on the road, Tesla can access to real-world data to help improve its FSD technology. Compared to Waymo, Tesla will have increasingly greater data advantages moving forward. Taken together, Tesla has the data, money, and leadership to make its no-lidar FSD a reality.
Forward Outlook: Tesla’s Economic Moat
In the following, I discuss why technological advantage (FSD) and branding power can help Tesla build a strong economic moat.
Competitive advantage 1: Leading in FSD
Electric cars and FSD are two concurrent but parallel technologies in the auto industry. Major auto manufacturers recognize the importance of electric cars and have started to invest in it. For instance, in 2019 European carmaker Audi announced an acceleration of its electric vehicle plans with an investment of €12 billion from 2020 to 2024. Similarly, Ford in 2018 announced to invest $11 billion by 2020 on electric cars. For electric cars, battery range is a key indicator for competitive advantages. Tesla is leading in battery range, such that there a 132 mile gap between Tesla Model S and the highest non-Tesla electric car on the market (see Figure 8).
(Figure 8: Tesla leading in battery range; Source: Tesla 2020 Q1 Conference call)
Different from electric cars, FSD is more about computer programming and statistical theories. That is, FSD is to use programs and algorithms to understand images captured by cameras, along with information from other sensors, to make acceptable diving decisions. FSD programs and algorithms are still developing and evolving at a rapid pace in both academia and industry (Side Note: Depending upon the context, such algorithms and computer programs may be referred as neural network or machine learning, terms to describe that they can utilize external information to make decisions like human beings.)
(Figure 9: Tesla to offer the function of recognizing and responding to traffic lights and stop signs; Source: Teslarati)
(Figure 10: Tesla FSD hardware, including chips and integrated circuit board; Source: Theverge.com)
I believe that Tesla can leverage FSD to build a strong economic moat for two reasons. First, it is challenging for traditional carmakers to build a strong in-house FSD R&D team. As mentioned above, this is due to the fact that FSD builds on the expertise that is very different from traditional car engineering. It thus partially explains why three major carmakers in the US either acquired or started a partnership with specialized FSD developers (see Table 1). Tesla is leading in FSD such that it plans to offer the function of recognizing and responding to traffic lights and stop signs by the end of this year (see Figure 9). For FSD, major carmakers neither started as early, nor invested as heavily, as Tesla. Consequently, Tesla has significant first-mover advantages in FSD.
Second, Tesla has made significant progress in FSD. FSD involves two parts: (1) hardware: the chips and the integrated circuit board (see Figure 10) (2) software: the programs (or, algorithms) running on the hardware to make driving decisions. Tesla started to design FSD chips in 2016, and to use them in Models S, X, and 3 since 2019. According to Tesla, such hardware can be upgraded to FSD simply via software update, even though to date it is only running autopilot. Tesla plans to release the FSD software by the end of this year or next year.
I believe Tesla’s progress in FSD is significant in two aspects. (1) Tesla is the only developer designing FSD chips in-house (see Table 1). Other FSD developers rely on either Intel (INTC) or Nvidia (NVDA). Further, Tesla is the only carmaker selling FSD hardware to individual customers, suggesting that Tesla is confident in its FSD hardware and upcoming software. Thus, the transition from Tesla autopilot to FSD will be smooth, as Tesla could just release more and more features moving forward. Such a transition will be continuous and progressive (similar to Apple updating iOS) rather than abrupt and binary (without FSD vs. with FSD). (2) Tesla has better control over developing FSD moving forward. Again, I use Apple to illustrate this. In particular, Tesla’s integration of FSD and car manufacturing is very similar to the approach of Apple. Apple designs its iPhone hardware; more importantly, different from Samsung using Google Android, Apple designs its operating system iOS as well. As we know, iOS gives Apple great advantages when competing with Samsung in the smartphone market. Taken together, I believe that Tesla meets Buffett’s criterion of “technological advantage.” Tesla’s in-house FSD approach can help it achieve an economic moat in the long-term.
Competitive advantage 2: Strong branding power.
(Figure 11: Apple’s advertising campaign; Source: Business Insider)
They (consumers) are more interested in Tesla, more so than looking for an electric car.
– Jake Fisher (Senior Director of Auto Testing at Consumer Reports) in a CNBC interview
The observation of “A Tesla vs. An electric car” reveals that consumers do not just view Tesla as one of the many electric carmakers, but rather an electric car with a distinct brand identity. This is very similar to Apple’s campaign of “PC vs. Mac,” which is a textbook example of product positioning and branding. While we typically do not view electronic products as luxury or status goods, Apple is considered as a premium brand by the broad public. This is a bit different from the traditional definition of luxuriousness, which typically associates with Swiss watches or Italian handbags. Such nuance partially explains the disconnection between analysts and market realities. In 2017, a Fortune article wrote “Apple’s Reported $1,000 iPhone 8 Price Is Putting Off China’s Shoppers” as the price was roughly double the average Chinese monthly salary. We know it turns out to be incorrect, and a significant portion of Chinese are willing to pay for a $1000 iPhone. Based on Bain, China will account for roughly half of the global luxury market by 2025. With the new factory in Shanghai, Tesla can leverage its branding power to sell cars at a premium, contributing to its growth.
Chinese shoppers ramp up their purchasing, especially in China. By 2025, Chinese consumers will account for 46% of the global (luxury) market (up from 33% in 2018), and they will make half of their purchases at home in China (up from 24% in 2017).
– Bain (Luxury Goods Worldwide Market Study)
(Figure 12: Elon Musk’s tweets; Source: Twitter)
How does Tesla achieve such strong branding power? First, Tesla is the first carmaker to solely make and sell electric cars to the general public. The pioneering role helps Tesla build a positive brand image of innovation. Tesla’s leading role in battery, product design, autopilot, and FSD further strengthens the brand image. Second, it can be attributed to Elon Musk, who acts as the spokesperson and promoter for Tesla. Marketing research reveals founders are an integrated part of brand identity, such that Apple is associated with Steve Jobs. In the same vein, Elon Musk’s entrepreneurial spirit in Tesla and SpaceX helps build the brand of Tesla. In theory, Tesla and SpaceX are two independent entities. Yet, the connection via Elon Musk nudges consumers into associating Tesla with SpaceX. From Figure 12, we see Musk shared exciting news about SpaceX as well as promoted Tesla cars via his Twitter (TWTR) account. Recall Falcon Heavy sent a Tesla Roadster into space in 2018, which essentially was a free commercial for Tesla. To summarize, Tesla’s strong branding power is consistent with Buffett’s definition of “its position in the consumer’s mind.”
Conclusion
Although Buffett does not buy Tesla stocks, Tesla meets Buffett’s criteria for a strong economic moat. In particular, Tesla is leading in the technology of battery and FSD. Importantly, Tesla is able to keep investing in R&D, which will strengthen and enlarge its technological advantages moving forward. Further, Tesla has an enthusiastic fan base, and such customer loyalty empowers Tesla to sell cars at a premium in the United States and international markets. In my opinion, accordingly, Tesla is a buy due to its strong economic moat.