Text: Internet Jianghu
Mobileye’s market cap collapsed, but autonomous driving didn’t.
No one expected that the valuation of Mobileye, an autonomous driving technology company owned by Intel, collapsed.
According to Intel’s public filings, Mobileye’s IPO was valued at $15.9 billion. A far cry from the $50 billion valuation at the end of last year, the valuation has shrunk by more than $30 billion in just a few months.
According to the market value of Xiaomi Group’s US$28.4 billion on the latest trading day of Hong Kong stocks, Mobile’s valuation has fallen by one Xiaomi…
Compared with Mobileye being poured cold water, the domestic autonomous driving track is still full of capital. According to incomplete statistics, a total of 44 autonomous driving companies at home and abroad completed financing in the third quarter, with a total financing amount of about 1.33 billion US dollars, and a total of 4 companies received more than 100 million US dollars in financing.
In addition, some startups in the autonomous driving segment have even received money. According to the financing information of the Tianyancha APP, a cross-vehicle cloud data-driven empowerment company called Zhixiehuitong received tens of millions in the third quarter. A round of financing in RMB.
Some analysts believe that one of the reasons why the market is no longer optimistic about Mobileye is that it is difficult for the Mobileye autonomous driving chips that have been put into the market to maintain market competitiveness. The competitiveness of NVIDIA, Horizon, Huawei, Baidu and other major manufacturers is already very strong. Manufacturers have also begun to abandon Mobileye’s products and solutions.
Mobileye’s valuation has retreated sharply, and there are certain warnings for the autonomous driving track. Although capital enthusiasm is still there, as autonomous driving has come to the eve of scale, the value of autonomous driving is being consolidated layer by layer. Autonomous driving companies must not only win the “pole position” to win the starting line of financing, but also always stand in the front row.
Behind the “diving” of Mobileye’s valuation, how far is the commercial future of autonomous driving?
Mobileye is already a very leading company in the industry. According to the disclosed prospectus information, from 2019 to 2021, Mobileye’s revenue will be 879 million US dollars, 967 million US dollars and 1.4 billion US dollars respectively.
In other words, in the past three years, Mobileye’s revenue has maintained a stable growth rate of more than 10%. Compared with the half-year revenue as of July 2, 2022 and June 26, 2021, the year-on-year growth has reached 20%. .
The reason why the capital market does not buy it may lie in the iterative rhythm of product technology. From the product point of view, Mobileye’s products are mainly to help solve the realization of vehicle collision warning, AEB emergency braking, ACC adaptive cruise and other capabilities.
And a number of terminal manufacturers including Weilai, Tesla, and Ideal have begun to abandon Mobileye. The ideal choice is Horizon, Weilai chose NVIDIA, and Tesla used its own FSD. Although the next generation of ADAS is under development, the iterative pace has not kept pace with the development of autonomous driving.
After going through the initial stage of research and development, once autonomous driving enters the landing stage, it has to pay attention to changes in market demand.
Mobileye’s lessons have also brought a wake-up call to some companies that were originally doing L4 but turned to L2 autonomous driving. Even if the low-level assisted driving is “milk bread”, it still needs to look forward at all times and keep up with the pace and rhythm of the industry as a whole. .
From Mobileye’s experience, the current commercialization of autonomous driving should pay more attention to the issue of iteration.
One of the most basic principles in making commercial products is that products and solutions need to keep up with iterative market demands.
For Internet car builders, this is all too familiar. Isn’t it the version iteration in Internet products. But in the automotive world, iteration may not be so easy.
On the one hand, for autonomous driving companies that empower OEMs, can the cost increase brought about by the iteration of solutions be accepted by car manufacturers? On the other hand, in the implementation stage, the iteration of the solution is not only the iteration of the software. Does the supplier need to adjust it? These issues need to be considered, so it will be more difficult to grasp the iterative pace of product solutions for autonomous driving.
The key to solving this problem is to find a balance between short-term commercialization and long-term technological competitiveness with reasonable resource allocation.
For start-up companies, it is no longer realistic to pursue technological leadership in one step. How to take into account the landing of L2 and the research and development of L4 and L5 technologies is not only a strategic issue, but also a survival issue. After all, the L4 tests of leading players such as Baidu Apollo, Xiaopeng, and Weilai are already underway, and there is not much time left for autonomous driving companies.
In addition to finding a balance point, another key to the commercialization of autonomous driving is cost control.
At present, there are three main routes for autonomous driving in the world: pure vision, multi-sensor fusion and full fusion algorithms. It is difficult to say which of these three schemes has an absolute advantage. However, whether it is the battle of algorithms or the battle of radar and vision, one issue that cannot be avoided in the end is mass production.
Tesla has proven to the industry that mass production is the most cost-effective way to improve self-driving technology. Low-cost large-scale applications will accelerate the dawn of the real commercialization of autonomous driving.
Whether it is an OEM, an autonomous driving company, or a major manufacturer such as Baidu and Huawei, all of them are seeking mass production of autonomous driving. Baidu has gone a long way in Robotaxi, and Huawei also aims to assist driving by asking the world.
However, the purpose of mass production of autonomous driving is not only to improve the technical level, but also to take the lead in reaching the commanding heights of scale in terms of commercialization.
Whether it can achieve L5 at the current stage is not really important, the important thing is to seize the market first. Such a business strategy has played out countless times in the Internet world, and it is likely to be staged in the automotive industry as well.
However, to seize the market, the consumer side cannot be ignored. Internet Jianghu believes that autonomous driving technology is essentially not just a ToB technology product, but a ToC consumer product.
When will autonomous driving cease to be self-inflicted by the industry and truly enter the public eye? Probably when the Model 3 is mass-produced.
That is to say, Tesla’s biggest impact on the autonomous driving industry is not actually how advanced its FSD is, but that after a large number of middle-class people bought the Model 3 for more than 200,000 yuan, they had a concept of the technology of autonomous driving.
If the cost of a technology cannot be reduced to a price that can be afforded by mass consumption, then the technology is worthless to the capital market. Only when people accept that autonomous driving is a real “consumer product” can we say that autonomous driving is a thing!
That is to say, when the consumer market truly recognizes and accepts it, the commercialization of autonomous driving will really bear fruit, and the capital market will also give a better valuation.
Autonomous driving in the context of commercialization, has the bicycle intelligent pie really won?
The maturity of technology development ultimately determines whether it can be successfully commercialized, but conversely, the most advanced technology under commercialization requirements is often not the most suitable for the needs of the market.
The landing of autonomous driving has been called for a few years, but until today, the battle between visual routes and lidar fusion routes, the battle between bicycle intelligence and vehicle-road coordination, and the battle between urban, trunk, wharf, and mine landing scenes still have no definite results.
Han Xu, founder and CEO of WeRide, believes that bicycle intelligence is a headlight, and vehicle-road coordination is a street light. The former must exist, but the latter is not.
In “Intelligent Transportation”, Robin Li’s point of view is more clear. He believes that autonomous driving is the starting point and the end is intelligent transportation. Robin Li also did not shy away from telling the outside world that the use of vehicle-road coordination solutions for autonomous driving is a technical route that Baidu insists on and is optimistic about.
From a commercial point of view, Bicycle Intelligence seems to be a solution that can be effectively implemented.
Taking the Ideal L9 as an example, the BOM cost of a complete set of high-end intelligent driving can reach tens of thousands of yuan. Qingzhou Zhihang has also released a 10,000-yuan vehicle-level autonomous driving solution. That is to say, there is still room for further reduction in the cost of realizing bicycle intelligence.
Bicycle intelligence brings us functions such as automatic parking and adaptive cruise in intelligent driving, and the road to commercialization in the early stage will be better.
The vehicle-road coordination is an issue that the industry needs to consider from the 10th to 100th stage. Starting from the overall system efficiency, vehicle-road collaboration must be the best smart transportation solution.
In fact, in terms of landing, the markets faced by the startups of the two routes are not exactly the same. In other words, the competitive pressure of the two routes does not come from commercialization and market competition.
Companies that do bicycle intelligence, such as Waymo, Cruise and Mobileye, their customers are mainly OEMs and car brands.
As an autonomous driving company with vehicle-road coordination, it is more interested in the ToG market. Because the construction of intelligent roads requires 5G, and the intelligent transformation of existing roads is required, all of which need to be promoted by the G side. Therefore, the biggest problem with the implementation of vehicle-road coordination is cost. It is reported that the current cost of a one kilometer smart road is about one million yuan, which has not yet reached a cost range that can be truly scaled.
Internet Jianghu believes that the real competition may occur in the capital market.
After 2019, the frequency of industry financing began to decrease, the industry entered the involution, the easiest stage of financing has passed, and autonomous driving companies need to return to blood while maintaining their operations.
The domestic self-driving unicorn financing situation was not very optimistic. For example, Momenta has not received new financing for more than two years since the B+ round of financing in October 2018. Until September 2021, some media reported that it had obtained a financing from GM. 400 million US dollars investment.
In fact, after 2019, most of the capital that invested in autonomous driving companies also invested in the industry chain, and some traditional car manufacturers will strengthen their future strategic layout by injecting capital.
In China, the capital market is actually watching.
On the one hand, after getting accustomed to the high investment and fast return of the consumer Internet, it is not easy to turn to the field of investment in technology. On the other hand, after the first wave of capital entered the field of autonomous driving in 2018 and 2019, the original investors The return window has not yet appeared, and the difficulty of AI commercialization makes the capital market still have doubts about autonomous driving.
Therefore, in the autonomous driving track, the capital market faces a multiple-choice question, whether to invest in “vehicle-road coordination” or “bike intelligence”.
Behind this question, there are actually two choices of value logic.
The value chain of bicycle intelligence is actually an industrial value chain. Head autonomous driving companies, smart car manufacturers, and head autonomous driving chip companies are obviously more valuable.
In other words, this is actually a long-term value investment path, and the return cycle may be longer. Coupled with the risk of changes in the future market competition pattern, the prospect is not so clear. However, the automobile industry chain involved in bicycle intelligence is long enough and large enough, which means that once the scale of autonomous driving is mature, the value return may be higher.
The value chain of vehicle-road coordination is actually more like an infrastructure value chain.
In the long run, even if bicycle intelligence has reached a very high level, under the guidance of governance, vehicle-road coordination must be the overall best solution. In addition, the investment scale of infrastructure construction in the direction of smart transportation is also in the trillion level, and the policy direction is clear enough, but it may not be so easy to commercialize it on a large scale in the short term.
Write at the end:
The world returns in the same way but has different paths, and it is the same and has a hundred concerns.
Autonomous driving, will eventually embark on the same road to the future.
Technology is not the purpose, the purpose is to make technology into products and provide better services for people. In the long run, the autonomous driving industry, where a hundred flowers bloom and a hundred schools of thought contend, may not be a technological stage play worthy of people’s expectations.
Who can dance to the end and win applause under this gorgeous light? Maybe time will give the answer.