Can Mobileye, the "godfather" of autonomous driving, make a comeback after joining hands with Geely?

The technology market will only have a dominant player in its infancy. Once a hundred schools of thought contend, it actually means that it has entered a golden age.

The 2022 Beijing Winter Olympics has not yet opened, but Mobileye, once the "leader" in the autonomous driving industry, has already made a big splash in the spotlight.

On January 24, Intel demonstrated many of its new technologies that will be used in the Beijing Winter Olympics. Among them, the safe travel assurance system from its autonomous driving technology company Mobileye was particularly eye-catching.

According to media reports, the Beijing Winter Olympics Organizing Committee and the Beijing Transportation Bureau have approved a total of 500 vehicles equipped with autonomous driving equipment from multiple companies to be used in venues and personnel transportation services, including buses equipped with Mobileye 8 Connect™ and Mobileye Shield+ collision avoidance systems.

At that time, all these smart vehicles will work together to provide travel services for all participating athletes, officials from various countries and invited media during the Winter Olympics.

“The vehicle is equipped with an artificial intelligence safety system that uses computer vision technology to assist driving. The system continuously scans the road ahead, and once a pedestrian or cyclist enters the danger zone, the system will issue a visual and audible alarm in real time to remind the driver to take action.”

▲Mobileye 8 Connect™ Intelligent Collision Avoidance System

In fact, for a company like Mobileye that insists on pure vision-based intelligent driving, its intelligent driving system is still challenging when facing scenes such as snowy roads in winter. What's more, Beijing is a city that is well-known for its traffic problems.

Therefore, being able to put it into use during the Beijing Winter Olympics is both an opportunity and a challenge for Mobileye.

If the company succeeds, it will further become famous, not only will the company be able to reverse the downward trend since the beginning of 2020, but it will also have the opportunity to get a big order from the Chinese government. However, if a major accident occurs, the future of this company may be quite dark.

However, this is a risk that both Mobileye and Intel have to take. After all, it is only the first month of 2022, and Mobileye has announced several major events in succession, reaching cooperation agreements with Volkswagen, Ford and Geely Auto (ZEEKR brand).

Its appearance at the Beijing Winter Olympics also means that Mobileye, which once dominated the smart driving market, has made up its mind to turn the tables at the beginning of 2022 after experiencing the painful cross-border squeeze from "graphics card manufacturer" Nvidia for two years.

01. The former king

At the end of 2021, Intel announced that it would promote Mobileye, its smart driving subsidiary, to be listed independently on the New York Stock Exchange. At that time, the market valuation of Mobileye was as high as US$50 billion.

In the field of intelligent driving, Mobileye, founded in Israel in 1999, can be regarded as a veteran company. The outside world can understand its specific business just from its corporate name, "the eyes of the car".

Mobileye started out as a visual algorithm company. Its main business is developing driver-assisted systems and advanced driver assistance systems (ADAS) for automakers. At the beginning of this century, Mobileye took the lead in launching the world's first smart driving chip, EyeQ 1. Subsequently, EyeQ went through six iterations and developed into a whole series.

In the field of intelligent driving, it is not an exaggeration to call Mobileye a "veteran". To some extent, it may not be an exaggeration to call it the "ancestor". After all, many standards and basic functions in the current vehicle intelligent driving system, such as the front collision warning, lane departure warning, and blind spot detection functions that are standard in the current mainstream ADAS systems, were first developed by Mobileye and the standards were defined.

From the founding of the company to 2019, Mobileye has monopolized the entire smart driving market for a time by using algorithms + EyeQ series chip solutions, thanks to its forward-looking approach in seizing the first wave of the integration of automobiles and artificial intelligence, helping many automakers realize the initial stage of smart driving functions.

Until 2018, Mobileye actually enjoyed a de facto monopoly in the global smart driving market. And this de facto monopoly continued until Nvidia, the "graphics card seller", decided to enter the autonomous driving industry.

In 2020, Nvidia launched Orin, a dedicated chip for autonomous driving, which immediately caused a "strong earthquake" in the industry.

The shock first comes from the performance. The Orin chip has excellent performance, with a single chip computing power of up to 254TOPS, and multiple chips can be used together to further enhance the computing power of the system. In addition, NVIDIA has mastered many unique skills in graphics processing due to its long-term "graphics card manufacturing".

Obviously, once such a company decides to cross over to the autonomous driving business, it will inevitably have a major impact on the original industry structure. At one time, Weilai, Xiaopeng, Ideal, WM Motor, Jidu, Mercedes-Benz, Volvo, and the Zhiji series under SAIC Group announced that they would adopt NVIDIA products. A considerable number of them had been partners of Mobileye for many years or even more than ten years.

What’s worse, Nvidia is not resting on its laurels after Orin’s success.

In 2021, "Leather Jacket Old Huang" continued his efforts and launched the next-generation product Atlan with 1000TOPS computing power, and plans to launch it on the market in 2023.

Of course, Nvidia is not the only company that poses a threat to Mobileye. In fact, Chinese technology companies including Black Sesame, Horizon Robotics, and later Huawei have also launched their own powerful smart driving chip products and supporting solutions between 2019 and 2021. Although the threat from these companies is not as great as Nvidia, they have indeed taken away a lot of market share from Mobileye.

Now, let's go back to Mobileye's independent listing plan. Because of the above "background", after Intel announced its listing plan, there were many voices in the market that were pessimistic.

Many people believe that the end of Mobileye's cooperation with many car companies means that its former position as the leader in autonomous driving has been substantially shaken. And it seems almost impossible to restore its former status.

The reason why it is "impossible" is partly due to the performance shortcomings of Mobileye's hardware products.

For example, EyeQ 5, which was taped out last year, achieved ultra-low power consumption of only 10W with its 7nm process, which is top-notch performance. However, in terms of computing power, EyeQ 5 only has 24TOPS, which ranks last among the new generation of dedicated chips for autonomous driving.

▲Although the "computing power only theory" is unnecessary, it is really too much to have only 24TOPS until 2021

Although there are many schools of thought in autonomous driving, the "heavy perception" route that relies on pure vision is known for its low requirements for system computing power, and EyeQ 5 is sufficient for now. However, too low performance will inevitably affect mid- and long-term upgrade issues.

In addition, Mobileye's long-standing strategy of integrated hardware and software solutions is becoming a serious obstacle to the company's continued development.

Since the software system that Mobileye provides to its cooperating OEMs is "closed", this means that the automakers cannot modify and adjust its algorithms independently, which limits their autonomy. In contrast, companies including Nvidia, Black Sesame, Horizon Robotics, and even Huawei provide customized algorithm adjustments.

Although Mobileye's intention to maximize its profits by controlling the entire autonomous driving domain is understandable, this will inevitably make OEMs more inclined to consider partners that give it more autonomy when there are other options. There was no choice before, but now that there are so many options, why should we endure harsh conditions and insist on "hanging on a tree"?

02. Mobileye’s “transformation”

When faced with survival pressure, most companies will make wise choices. Mobileye is fully aware of the crisis it faces and is willing to make changes.

At the 2021 CES held at the end of last year, Mobileye and Geely's Zeekr brand jointly announced that the two parties will cooperate to develop pure electric smart cars with L4 autonomous driving capabilities.

According to the official announcement from both parties during CES, the two companies will adopt a highly divided approach in the development of the new car. The vehicle body will be built based on the SEA architecture launched by Geely in September 2020, while Mobileye will be responsible for the development and integration of its intelligent driving system. It is reported that the EyeQ 5 chip will still play a leading role in this new car, and up to 6 chips will be equipped.

The cooperation project with Geely marks the first time that Mobileye has been deeply involved in a "car manufacturing" project. Previously, the company had long insisted on only providing autonomous driving solutions to OEMs.

In addition to personally participating in the "car manufacturing" plan, another major change Mobileye has shown is giving up its complete monopoly on the solution in its cooperation with Ford Motor Company.

New energy and intelligent transformation are currently Ford's top priorities. Since it is obviously a step behind European and Chinese automakers, it is imperative for Ford to develop a set of assisted driving systems that can solve the intelligent transformation of its various models as soon as possible.