In 20 17, Intel acquired Mobileye at a sky-high price of153 billion, which was one of the biggest acquisitions in Intel's history. In the days when it was acquired, Moblileye, Intel's "own son", can be described as "entrusting people to give money". It not only assigned its own laser radar and 4D radar development team to Moblileye, but also spent $900 million in 2020 to acquire Moovit, an Israeli urban traffic data startup, to help Mobileye quickly overcome its shortcomings and become a "full-stack self-research" giant in the new era of autonomous driving.
Mobileye also lived up to expectations, not only becoming the new growth pole of Intel with its bright financial performance, but also firmly sitting on the throne of "King of ADAS" with nearly 80% market share.
In terms of financial data, the operating income of mobile ye 202 1 in the fourth quarter was US$ 356 million, up 7% year-on-year, continuing to set a new record. The single-season revenue data is comparable to the annual revenue of $358 million in the fourth quarter of 20 16. At the same time, the company's revenue in 20021year reached1400 million USD, up 43% year-on-year, and the compound growth rate from 201year to 20021year was as high as 26. 1%.
In terms of sales volume, Moblieye ushered in a historic moment in June 5438+February 65438+ March, and the cumulative shipment of Mobileyeye chips exceeded 65438+billion! Since the mass production of EyeQ 1 chips in BMW, GM and Volvo in 2007, the sales of EyeQ series chips have been increasing continuously. From 20 18 to 202 1, the sales data were124,000,175,000,193,000 and 28/kloc-0,000, with year-on-year growth rates of 43% and 4/0,000 respectively.
In addition, in 20021year, Mobileye also received 4 1 new orders from more than 30 car companies, involving about 50 million new cars.
Thanks to the strong financial performance, Intel decided to let Mobileye sprint IPO as soon as possible, despite the recent sluggish performance of the US stock market due to geopolitics and interest rate hike expectations.
Wall Street mentioned that on March 7, Intel secretly submitted the registration documents to the US Securities and Exchange Commission, and planned to issue Mobileye new shares for the first time. The market expects its valuation to exceed $50 billion, which will impact the largest IPO in the history of autonomous driving.
However, since Mobileye is in a good situation and the market prospect of autonomous driving is so broad, why doesn't Intel "hold" Mobileye for a few more years and then hit a larger IPO when the market is "crazy" again?
At the end of the 20th century, Advanced Driver Assistance System (ADAS) was still in its infancy, and the whole industry believed that only expensive radar sensors could realize the necessary functions such as automatic emergency braking (AEB).
However, Amnon Shashua, a professor of computer sustainability at Hebrew University, proved in his research that basically all the perceptual tasks necessary for driving assistance can be realized based on a single camera. This makes the cost of driver assistance system fall to "civilian level" in theory, thus making it possible to commercialize it.
With vigorous skills and great courage, Uncle Sha started his business with two friends and set up Mobieye Company in 1999 * *, focusing on ADAS circuits. As the name implies, the company intends to use computer vision technology to detect vehicles, become the "eyes" of vehicles, and prevent collisions to reduce vehicle accidents and casualties.
200 1, constrained by hardware, Mobieye realized that "people who really write software seriously should make their own hardware". The company's leadership believes that only by designing its own Soc and solving the heavy load caused by computer vision can we give full play to the company's full strength. Great minds think alike, and Mobileye then embarked on the same road of software and hardware integration as Apple.
In 2004, under the leadership of Elchanan Rushinek, Mobileye's SoC design team successfully developed Eyeq 180nm process.
On the basis of "internal and external repair" of software and hardware, Mobileye successfully combined the leading computer vision algorithm, powerful EyeQ Soc chip and cheap single camera to create a cost-effective monocular vision ADAS, which was welcomed by Volvo, BMW, GM, Audi and other major automobile manufacturers.
In 2007, with the investment of $654.38+300 million from Goldman Sachs, Mobileye gradually grew into an "autopilot" unicorn that made investors crazy, especially after the launch of EyeQ3 in 2065.438+04. In the same year, Mobileye successfully landed on the New York Stock Exchange, with a market value of 8 billion US dollars. At the same time, it set a record for the largest IPO of Israeli companies in the United States, raising 890 million US dollars.
Seeing the broad prospects of Mobileye, in 20 17, Intel, the "blue giant" who intends to seize the highland of autonomous driving technology, pocketed Moblieye at a high price of about/kloc-0.5 billion US dollars, with a premium rate as high as 34%. Since then, Mobileye has become Intel's autonomous driving business unit. After completing the transaction, Brian Coe, then CEO of Intel, said excitedly:
After joining hands with semiconductor giant Intel, Mobileye continues to be invincible in the field of autonomous driving chips. Before 2020, smart driving chips were almost monopolized by Mobileye. According to the latest data from research institute Guidehouse Insights, Mobileye accounts for about 80% of the global market for advanced autopilot-assisted vision systems.
However, Mobileye, who is proud of the spring breeze, may have a hard time in the future.
Automatic driving is usually divided into six levels: L0~L5, from manual driving, assisted driving to conditional automation, and finally fully automatic, with increasing complexity.
In the past five years, driven by Tesla's demonstration, global car companies have seen the potential of autonomous driving, and major car manufacturers are interested in launching autonomous driving products, and autonomous driving is moving towards L4/L5 high speed.
According to iResearch data, after 2020, models with L3 function such as TJP and HWP will gradually land; The landing of L4/5 is expected to be no earlier than 2025, and the scene will gradually develop from a closed driving environment (parks, ports, mines and airports) to an open road scene in the city. By then, the vehicle will be completely taken over by the system, the attributes of vehicle travel tools will be strengthened, and the way consumers buy cars and travel will be greatly changed.
The trend of automatic driving to L4/L5 highway has brought great challenges to Mobileye, the former king.
The computing power of (1) chip lags behind, and EyeQ5 is "untimely".
As mentioned above, Mobileye's main first-Mover advantage is based on "reducing computing power and reducing costs". Therefore, in the early days of L 1, L2 and other autonomous driving, there is little demand for computing power, but Mobileye, which integrates software and hardware, is invincible in the world and its market position is hard to shake.
With the new functional requirements brought about by the upgrading of autonomous driving level, millimeter wave radar, ultrasonic radar and emerging laser radar have appeared in the next generation models of various OEMs in addition to the traditional cameras. The increase of sensors will also bring huge data processing pressure, which requires chips with higher computing power to process, analyze and make decisions. Therefore, the higher the computing power, the more complex situations that can be handled, and the safer the automatic driving (assisted driving).
As far as chip computing power is concerned, Mobileye's EyeQ series Soc is somewhat outdated. Before 202 1 year, Mobileye's most advanced chip was mass-produced in 20 18, which only supported L3 autopilot EyeQ4. EyeQ5 supporting L4 came late, it was not listed until 2020, and it was delivered until 202 1. At the same time, EyeQ5 has only 24TOPS computing power and 10W power consumption, and its performance will be overshadowed by "rising stars" such as NVIDIA, AMD, Huawei, Qualcomm and Horizon.
In particular, NVIDIA, which has a GPU resource endowment and a full-stack tool chain, has more market prospects in the field of advanced autonomous driving. Team Wu of Guotai Junan Securities said:
The market also confirms this. Today, Weilai, Tucki and Ideality are all customers of NVIDIA, and traditional OEM manufacturers such as Chi Machine Automobile and SAIC R are among them. In addition, traditional car companies such as Volvo, Mercedes-Benz and Buick, and commercial vehicle manufacturers such as Didi and ZOOX are also customers of NVIDIA.
It is worth mentioning that domestic manufacturers are not to be outdone in computing power, keeping up with the international leading level. For example, now Huawei's MDC series platform, with computing power of 400+TOPS, has been mass-produced, while the new generation chips of Horizon and Black Sesame, with computing power of several hundred TOPS, have been lost.
(2) too closed, a large number of customers change doors.
Mobileye uses software and hardware packaging, and the "non-open source" approach also makes this autopilot "big brother" criticized.
Take EyeQ5 chip with "CPU+ASIC" architecture as an example. Although the power consumption of this chip is extremely low, its algorithm system is relatively closed, which is almost a black box for OEM and Tier 1, and it is almost impossible to differentiate its own algorithm through two modifications. When autonomous driving is still in the "primary stage", a relatively reliable and complete scheme is undoubtedly the fastest and safest choice. However, under the background of the rise of "open" manufacturers such as NVIDIA, Horizon, Huawei and Baidu, this "closed" scheme of software and hardware integration that is difficult to tune is not so pleasing.
Therefore, a number of terminal manufacturers including Weilai, Tesla and Ideality have changed their doors. Tesla had a bad relationship with Mobileye earlier and replaced Mobileye with an FSD chip. Ideally, 202 1 year, EyeQ4 is also replaced by Horizon Journey 3; A group of car companies such as Weilai have chosen powerful NVIDIA Orin chips for the new generation of models.
In particular, BMW, which was written as a milestone by Mobileye, recently reached a cooperation with Ride of Qualcomm. According to the Wall Street News, a BMW spokesman said in June last year165438+1October that Qualcomm's chips will be used in BMW's "Neue Klasse" series cars, which is expected to be put into production in 2025.
In addition, because Mobileye has no R&D team in China, its R&D system still faces data supervision problems in China.
In September last year, the Ministry of Industry and Information Technology issued the Opinions on Strengthening the Access Management of Intelligent Networked Automobile Manufacturers and Products, which mentioned that personal information and important data collected and generated during the operation in People's Republic of China (PRC) should be stored in China in accordance with relevant laws and regulations. If it is necessary to provide data abroad, it shall pass the data exit security assessment. Obviously, the road data related to autonomous driving is obviously closely related to national security.
Generally speaking, Mobileye's first-Mover advantage is getting smaller and smaller in the future autonomous driving field of "computing power determines ability". At the same time, with the improvement of autonomous driving level and the embedded software and hardware of major manufacturers at home and abroad, the competition in this field will become more and more fierce.
In the face of the encirclement and interception of major "cutting-edge autonomous driving", Mobileye did not "serve the old". In 2022, the company is fully deploying advanced autonomous driving.
On the one hand, at the CES 2022 conference, the former "King of ADAS" released three advanced process chips with high computing power in one breath. Among them, EyeQ Ultra is an autonomous driving chip, which is based on 5nm process and has a computing power of176 tops; EyeQ6 Light and EyeQ6 High are both ADAS chips, manufactured at 7nm, with computing power of 5TOPS and 34TOPS respectively. Mobileye's commercial layout of assisted driving and autonomous driving "walking on two legs" is still advancing in an orderly manner.
On the other hand, Mobileye is also trying to deepen cooperation with local manufacturers and become "more open". As Wall Street has said, at the CES conference in early October of this year, at the influential consumer electronics show in Las Vegas, USA, Mobileye revealed that it is cooperating with Zeekr, a pure electric vehicle brand owned by Geely Automobile, and will launch the world's first car with L4 autonomous driving capability in 2024. At the same time, both parties will effectively integrate software technologies.
Although the future of Mobileye can still be expected, considering the increasingly "white-hot" competition pattern in the autonomous driving industry, the best time for its listing may be now.
As far as Intel is concerned, Wall Street mentioned that since 2020, under the attack of giants such as AMD and Apple, Intel has become more and more difficult. The total market share of CPU dropped from nearly 80% a year ago to less than 65%, and the stock price also fell to the level of two years ago, and the company fell into a "dark moment". Now the company is pinning its hopes on IDM2.0 strategy, expanding the scale of chip foundry and "reviving the glory", which is at a time when it needs a lot of money.
For Mobileye, Intel's poor performance in recent years can be described as "the city gate is on fire and the fish in the pool are affected." Mobileye, which hoped to "enjoy the cool with its back against the big tree", also performed poorly in the past two years when its parent company was in a trough, and its technological progress was far less than that of its competitors. Under the attack of "rising stars" such as NVIDIA, Qualcomm and Huawei, Mobileye's leading edge is getting smaller and smaller.
So after a few years, will Mobileye's performance be as bright as it is now?
Therefore, for Intel, with the increase of investors' interest in self-driving cars and Mobileye's eye-catching data and financial indicators, it is still necessary to carefully consider the IPO and seek a high valuation of more than 50 billion US dollars. Boosted by Mobileye's "high prosperity", Intel's depressed stock performance may be improved while gaining a lot of investment income.
(It is mentioned on Wall Street that in the past year, due to the concerns about Intel's financial situation, its share price almost fell sharply every time Intel announced its financial report. In the past year, Intel's share price has fallen by nearly 19%. )
It is worth mentioning that although Intel CEO Kiel Singh previously stressed that Intel does not need IPO funds to revive the "Blue Giant", Morgan Stanley analyst Harlan Sur still bluntly said in the research report: This move can alleviate the financial pressure brought by Intel's future capital increase and research and development. At the same time, Sur believes that Intel chose to let Mobileye go public at this time, and the timing is very correct. Sur said in the research report:
For Mobileye, choosing IPO will help it get rid of the low valuation brought by Intel and greatly enhance its corporate image, but it is also a beautiful thing. Perhaps as Intel CEO Kiel Singh said, more potential of Mobileye is expected to be further released after listing:
After the successful listing of Mobileye, the "King of ADAS" with sufficient bullets will undoubtedly further aggravate the "involution" in the field of autonomous driving. Historically, the "white-hot" competition of giants can often greatly promote the speed of industry progress and technological change, thus benefiting consumers more, which is undoubtedly loved by consumers.
Perhaps, the future has arrived, and the curtain of a new era of autonomous driving has been opened.