Five years ago, Ford Motor (F) promised to launch a car without a steering wheel by 2021. Five years later, people who want to buy this kind of car are still waiting for the vehicle to appear on the market. Facts have proved that the development of self-driving cars is a very difficult task. However, although the promise has not been fulfilled for a long time, the distance between fully self-driving cars and us may be closer than investors realize.
If you plan to take a taxi from Las Vegas McCarran International Airport and open the Lyft app, you will find that there is an option for self-driving taxis in addition to the regular taxi-hailing options. When the BMW 5 Series equipped with self-driving technology is parked in front of you, it may not look like it is said in science fiction, because there is a driver in the car and will chat with you like a taxi driver. But then you will notice some details of the vehicle: the display screen looks like the one in the 1982 science fiction movie “Tron” (Tron), showing people crossing the road, vehicles from all sides, and others in a pixelated manner. Street lights; the vehicle turns very smoothly; the steering wheel moves by itself. After getting off the bus at your destination, you will realize that you have just had a minimal life-changing experience.
Investors forget that the potential of autonomous vehicles (AVs) is forgiven. In December 2020, Uber sold its self-driving business. Compared with the high profile when it started, Uber was a lot low-key when it exited the business. Lyft (LYFT) sold its autonomous driving business unit to Toyota Motor (TM) this year. Even Tesla’s (TSLA) CEO Elon Musk’s promise of the imminent fully autonomous driving is beginning to sound a little hollow. Although the S&P 500 index has repeatedly set new highs, the stocks of companies dedicated to autonomous driving technology have fallen about 30% so far this year, and are down 50% from their 52-week highs.
However, autonomous driving is closer to reality than ever before. Lyft’s latest self-driving taxis are manufactured by Motional, a joint venture established by Aptiv (APTV) and Hyundai Motors (005380. Korea). In Las Vegas, some of these self-driving taxis are not equipped with safety drivers. Lyft plans to launch its self-driving cars in more cities by 2023. The fully autonomous vehicles launched by Alphabet (GOOGL)’s Waymo have already appeared on the roads of Arizona. General Motors (GM) is testing its self-driving taxis in San Francisco. With the support of capital, the auto industry has now embarked on a reliable path to push self-driving cars to the mass market. Investors also have many opportunities, but they have to be carefully explored.
At present, autonomous driving technology has reached level 4
Autonomous driving was first proposed by the Defense Advanced Research Projects Agency (DARPA) in 2004, and it has come a long way. SAE International, formerly known as the Society of Automotive Engineers, divides autonomous driving technology into 5 levels. Level 1 means that the automatic system can sometimes assist the driver to complete certain driving tasks. Level 2 means that the automatic system can complete certain driving tasks, but the driver needs to monitor the driving environment and be ready to take over at any time. At present, most car companies have Achieved this level. But when most people talk about true self-driving cars, they mainly refer to levels 3, 4, and 5, that is, the vehicle can complete all operations. Level 3 means that the driver will no longer need to stand by with hands and feet, and the machine can complete almost all driving operations independently, but the driver still needs to maintain concentration in order to deal with situations that may not be able to cope with artificial intelligence at any time. Both Level 4 and Level 5 autonomous driving technologies can be called fully autonomous driving technologies. At this level, the car can already perform all driving operations without the driver’s intervention at all, and the driver can also focus on other things. Aspects such as work or rest. But the difference between the two is that level 4 autonomous driving is suitable for some scenarios, usually in cities or on highways, while level 5 requires autonomous vehicles to be able to drive the vehicle completely in any scenario. Therefore, it can be said that level 5 is the real autonomous driving.
The autopilot technology of the Las Vegas airport taxi mentioned above belongs to Level 4. Lyft has completed more than 100,000 autopilots with the assistance of a safety driver. It is equipped with a driver because of these autopilots. Driving a car is a new product, and currently it can only be run on public roads, while casinos are private property. Soon, the taxi manufactured by Motional can travel to more than 3,500 destinations in Las Vegas without a driver.
The benefits are obvious. The vehicles are running very well, they will be signaled every time they change lanes, and they will not whizz around around vehicles waiting to turn right or left. It obeys speed limit regulations and travels smoothly. The taxi fare ranges from US$15 to US$35, which is the same as Lyft’s regular fare.
Such self-driving cars as Motional use a series of sensors to help them navigate the streets and lanes of the city. The cameras on the cars provide 360-degree panoramic monitoring of the road. The radar on the car is a small version of the radar used by the airport control tower. , It can play a role in bad weather and when the light changes. In addition, it is equipped with a lidar that is particularly good at observing distant objects. All information is entered into high-level software in the onboard computer, which can decode the data and make autonomous driving possible.
All these costs are high. The cost of Hyundai IONIQ manufactured by Motional, which Lyft plans to launch in cities outside of Las Vegas before 2023, may be as high as $150,000, which may exceed the budget of many people. Even if the price is more reasonable, checking and calibrating autopilot sensors and upgrading software will cost a lot of money. The current generation of self-driving cars are commercial vehicles, and their costs must be justified by earning income. They will directly compete with taxi companies, taxis, and 370,000 drivers in the U.S. with dedicated drivers. And this It’s just a small part of the approximately 270 million cars registered in the United States. But just as anti-lock braking systems and airbags have gradually penetrated into the automotive industry, automatic emergency braking and adaptive cruise control systems are becoming standard equipment in many cars, and more autonomous driving functions are coming soon.
Is Tesla FSD just a gimmick?
Don’t expect these to come from Tesla. Musk kept telling investors and his Twitter fans that Tesla is about to achieve fully autonomous driving, but this does not seem to be the case, at least not from SAE standards. Tesla’s car is equipped with a basic driver assistance function, Autopilot, with adaptive cruise control and lane keeping assist system. For an additional $10,000, drivers can upgrade to Tesla’s so-called “Fully Automated Driving” (FSD), which can actively guide the vehicle to and from the highway ramp, as well as automatic lane change assistance. , But FSD still only belongs to level 2 autonomous driving. Musk’s hyped fully self-driving cars may come out at the end of 2021. The autonomous driving technology of this type of car will reach Level 3, but even so, the driver still needs to sit in the driver’s seat and be ready to take over.
There is a reason for this. Tesla uses optical cameras to manage driver assistance functions, and as of May this year only uses optical cameras. Except for Musk, everyone insists that if you want to achieve level 4 autonomous driving, you must have lidar. The cost issue is considered to be one of the reasons why Musk only uses cameras. Depending on the technology, the cost of each sensor of lidar may be between US$500 and US$1,000, which is twice or three times the cost of the existing passenger car system with level 2 autonomous driving technology. Musk said on Tesla’s third-quarter 2020 earnings conference call: “Even if lidar is free, we will not install it.”
The problems facing Tesla’s investors and the company’s stock are expectations and reality. Mismatch between. Morgan Stanley analyst Adam Jonas has set a target price of $900 per share for Tesla stock, which is one-third of Tesla’s market value (i.e. 300 billion US dollars) from “network service” and “Tesla Mobility”, which respectively refer to FSD software sales and self-driving taxi business, although he believes that Tesla will not introduce level 4 or 5 autonomous driving before 2030 Technical car.
Cathie Wood of ARK Invest believes that Tesla’s self-driving taxis with Level 4 self-driving technology may be on the road in 2024 or 2025. This is also what she believes Tesla’s stock will reach One of the reasons why it can rise to US$3,000 per share in 2025 is that Tesla’s market value will reach about US$2.9 trillion based on this calculation. Uncertainty about the autonomous driving business is another issue that investors have to consider when looking at Tesla stock. The stock recently traded at $616.60 per share, which has lost about four cents since mid-January. The market value of one.
Lidar popularization will have to wait five to seven years
Musk’s view of lidar is correct, lidar is indeed very expensive. For this reason alone, the six lidar companies listed in 2020 and 2021 may find it difficult to achieve their sales targets. By 2025, the combined sales of Innoviz Technologies (INVZ), Luminar Technologies (LAZR), Ouster (OUST), Velodyne Lidar (VLDR), AEVA Technologies (AEVA) and AEye (CFAC) are expected to reach US$4.8 billion, but The expected sales difference between each company is very large. Ouster predicts that by 2025 sales will reach 1.6 billion US dollars, AEye only 290 million US dollars. Both companies are targeting similar customers and end markets, and both believe that they have leading technology and software, but they cannot be both right.
The six companies expect earnings before interest, taxes, depreciation and amortization (Ebitda) to reach 1.7 billion US dollars, based on this calculation, the average profit margin is about 34%. In contrast, the average profit margin of auto parts suppliers is less than 10%. New technologies can bring higher profits, because these are patented technologies, or because the importance of functions will reduce customers’ sensitivity to product pricing. But lidar does not seem to meet these two conditions. Lidar sensors will eventually become a commodity, and investors are actually betting on the integrated software of the two companies.
In addition, car companies are very sensitive to the pricing of special features. The cost of advanced safety systems that can achieve level 2 or higher autonomous driving may be as high as $1,200. If the cost has been so high, then car companies may not use lidar because each sensor costs so much money. History shows that the adoption of lidar may take some time. “It took us 15 years to apply radar to cars, and it took us several years to adopt optical cameras,” said Blair LaCorte, CEO of AEye. It will be applied to automobiles in seven years.”
The following four companies are making autonomous driving a reality
General Motors and Aptiv are big winners
General Motors, the major shareholder of the self-driving car company Cruise, and Aptiv, an auto parts supplier that cooperates with Hyundai Motor, will benefit from the trend toward autonomous driving.
Aptiv’s predecessor was Delphi Automotive Systems, which was spun off from General Motors in 1999 and went bankrupt in 2005. Former CEO Rodney O’Neal (Rodney O’Neal) started the company’s transformation from a low-margin and difficult parts supplier to a high-margin, high-growth business. The company now focuses on vehicle electrification and Autopilot software.
Today, Aptiv’s products and software functions can integrate all data from sensors everywhere in the car. The company can sell complete Level 2 autonomous driving systems to automakers, or it can sell only parts. Approximately 25% of Aptiv’s sales now come from “advanced safety systems”, which is an industry term for autonomous driving functions. In addition, the company also sells power and signal products. However, there is a natural synergy between these two businesses. The more complex the sensors and data of automobiles, the more complex power supplies they need.
In 1999, Aptiv became one of the first companies to install radar on cars. This system costs US$3,000 each, which is only the size of a tissue box. Today’s sensors are smaller, and the price is only tens of dollars.
The performance of Aptiv stock is very eye-catching. The return rate of the past five years is about 23%. Based on the estimated earnings per share of $5.10 in 2022, the stock market earnings ratio is 29 times, higher than 20 times of the S&P 500 index, and much higher. 11 times that of General Motors. However, it is expected that the average annual sales growth of the company in the next two years will exceed that of other large auto suppliers, the former is about 11%, and the latter is about 7%.
And these forecasts do not take into account the half of Aptiv’s investment in the joint venture Motional. Morgan Stanley’s Jonas said on Aptiv’s fourth-quarter 2020 earnings conference call: “If this can happen, the shares in the joint venture company…the part that your company holds, may be worth more than the entire Aptiv.” Jonas’s view may be correct. For example, after Microsoft (MSFT) invested in General Motors’ Cruise earlier this year, the latter is now worth $30 billion. Aptiv CEO Kevin Clark (Kevin Clark) said that Motional does not want to operate a fleet of taxis like Cruise, but that does not mean that the company cannot do so. Motional is a hidden asset that brings icing on the cake to a great business. According to Jonas’ calculations, Aptiv’s stock price may reach $205, which is 30% higher than its recent level.
General Motors, which owns the Cruise business, is another option for investors. The company holds approximately two-thirds of Cruise’s shares. This portion of the shares is valued at approximately US$20 billion, which is almost equivalent to a quarter of General Motors’ market value of US$87 billion. . Based on the expected earnings per share of $6.78 in 2022, GM’s automotive business has a P/E ratio of approximately 7 times.
However, most analysts are not optimistic about GM’s Cruise business. The profit of the autonomous driving business is only in the theoretical stage. At present, Cruise is consuming capital rather than making money. Despite this, 90% of analysts who research and analyze General Motors stock still give the company a “buy” rating, with an average target stock price of about $71 per share, which is about 15% higher than the recent level. With the inclusion of this part of Cruise, the target price may be raised to around US$85 per share, which is 41% higher than the closing price of US$60.08 per share on June 17.
This time, Alphabet, the parent company of Waymo, is not on the recommended stock list. Waymo has just completed a round of 2.5 billion US dollars of investment, and is operating its own car-hailing network, launched a self-driving taxi in Arizona, and plans to promote it in other cities in the next few years. Waymo may be more valuable than Cruise and Motional, but even with a valuation of up to $50 billion, Waymo only accounts for 3% of Alphabet’s market value. Alphabet invested in Waymo for a reason—Google used to be a technology company incubator—but Waymo is still too small to contribute to Alphabet.
The future belongs to self-driving cars, but self-driving cars are not the future of Alphabet’s business.