@ Tech New Knowledge Original
Author 丨 Sakuragi Editor 丨 Moonmi
Less than two years after joining the company, Wang Kai, CTO of Ideal Automobile, recently left his job. This is the departure of the highest executive since the establishment of Ideal Cars.
When Wang Kai first took office in September 2020, he was responsible for a total of 7 businesses, including the research and development and mass production of smart car-related technologies, electrical architecture, intelligent cockpit, automatic driving, platform development and Li OS real-time operating system. However, in the organizational restructuring at the end of January 2022, the businesses under Wang Kai's jurisdiction, such as autonomous driving, computing power platform and LiOS (operating system), have been handed over to Ma Donghui, co-founder and chief engineer of Ideal Automobile.
In just a dozen months, from the oath to the eclipse, it is not unspeakable.
Before joining Ideal Automotive, Wang Kai worked for Visteon, a global technology company, for 8 years, the highest as Visteon Global Chief Architect and Autonomous Driving Director, responsible for the planning and landing of the company's software and hardware product routes, is a world-renowned expert in the field of electrical and electronic architecture, domain controllers, automotive intelligent networking and automatic driving, and is also the core designer of smartCore, the world's first mass production vehicle-level cockpit center domain controller.
Coincidentally, in January this year, Huawei's self-driving president Su Zhen was also confirmed to leave the news. It is rumored that Volkswagen's acquisition of Huawei's autonomous driving involves billions of euros. Does the frequent turmoil in charge of autonomous driving from Tier1 to the main engine factory mean some kind of trend, or does the development of technology reach a critical stage of acceleration, or indicates that the dust has settled in a stage of autonomous driving?
The climax is also the beginning of the curtain
As the most special existence of the new car-making forces, the ideal has always emphasized its own positioning of "family", from the voice interaction of the whole car, to the convenient co-driver's entertainment of the triptych screen, from the care to the design of the elderly and children getting on the car, to the larger space in the back row, the ideal considers all the needs of family members. However, behind the emphasis on humanized experience, it was once criticized for its lack of intelligence.
Li wanted to say after the first quarter of 2021 financial report: "In the field of automatic driving, the company still has to make up lessons compared with Tesla, Xiaopeng and so on. At this stage, the goal of the ideal car is to continuously improve its own capabilities and deepen its own capabilities in software and hardware, including perception, decision-making, planning and other aspects. ”
After finding Wang Kai, Li thought of the second way is to increase investment, he said on the conference call, in 2021 the ideal enterprise is expected to invest in research and development will grow to 3 billion yuan, compared with 1.1 billion yuan in 2020 double growth. In the next few years, the investment in the research and development of the ideal car will continue to grow.
The reason for this planning, Li xiang said that there are two reasons: "One is that the challenges faced by the ideal car in the future will be more and more, and the other is that since the IPO last year, the company's cash reserves have become more abundant, and the core purpose of financing is to increase research and development." Li xiang said that the company plans to invest about tens of billions of yuan in research and development in the next few years, and it will be about 3 billion yuan this year.
For these investments, it is difficult to describe specifically which directions will be invested, which are generally divided into two dimensions, one is to invest in new models. The second is to invest in forward-looking technologies, and there will be a lot of overlap between the two.
At that time, the newly hired CTO was also full of confidence and announced his grand plan.
Wang Kai said that more money will be invested in research and development. In 2021, we will prepare for a new, large-scale, and upgradeable system architecture. The system architecture encompasses a new generation of advanced driver assistance systems, computer systems and our own automotive operating systems. Regarding the ADAS team, there are now more than 300 engineers in total. By the end of this year, the team had grown to more than 600 people.
However, to everyone's surprise, the ideal catch-up speed in the direction of intelligence, especially automatic driving, is extremely fast, and a smooth transition has been achieved.
According to the research report of Huatai Research Institute, on the ideal one2021 model, ideal officially abandoned the integrated solution provided by Mobileye, and turned to the intelligent driving chip equipped with Horizon Journey 3, with the advanced auxiliary algorithm independently developed, and greatly improved the intelligent driving hardware loading program, upgraded the ADAS camera pixels, increased the number of millimeter wave radar, and assisted the high-precision positioning, and continued to do a good job in hardware embedding for the later functional iteration.
In the future, Ideal will work with NVIDIA and Desay SV to further enhance the computing power of its products in intelligent driving systems with full-size SUVs released in 2022.
And the hero of all this seems to have little to do with CTO Wang Kai. Ideal Car people have revealed to the media that at the beginning of 2021, the intelligent driving of Ideal Car was switched from a supplier plan to a self-developed plan, when Lang Xianpeng made a "military order" for delivery, from the evaluation of Ideal Car employees, "the self-development plan was finally delivered well", but "Wang Kai's sense of participation is not strong". Since then, the ideal intelligent driving direction pattern has begun to stabilize.
The case of Huawei, it seems, may be another version of the ideal. In the industry consensus is getting stronger and stronger, the company model is gradually stabilized, and when there is no new story, it is a natural thing for the company to come out.
Huawei's problem is that no large OEMs are willing to become Huawei's puppets and give up control, the most obvious manifestation of which is SAIC's abandonment of Huawei's behavior, SAIC Chairman Chen Hong when asked whether he would consider working with third-party companies such as Huawei in autonomous driving, he said, "In this way, it has become the soul, SAIC has become the body." For such a result, SAIC is unacceptable and must take the soul into its own hands. ”
This has almost become the attitude of mainstream OEMs to Huawei, and at the same time, within Huawei, the attitude of cooperation with existing car companies is not without differences.
The former departing Huawei management once said: "Putting a Rolls-Royce engine on a tractor, it must still be a tractor, not a Rolls-Royce." "Behind the revealed seems to be Huawei's helplessness in cooperation with car companies."
Under the stalemate, we also see the dismal sales volume, and in the short term, Unless Huawei builds its own car, creates a complete adaptation system, and regains the model of the mobile phone era, it will not be able to change the status quo at all. When the internal structure is already very stable ideal, in the case of technical can not produce any big waves, what can be changed is the ratio of manpower and the company structure.
Subsequently, Su Zhen left his job and rumors spread that Volkswagen had acquired Huawei's autonomous driving business.
Whether it is Huawei or ideal, about automatic driving, whether it is tier1 or main engine factory, behind it there is a total demand for stable mass production and control costs, superimposed multiple considerations, and the heavy bet on automatic driving has obviously lost its cost performance.
Drive yourself in the water until there are no fish
Why is the pattern of automatic driving a clear card, first of all, from the structural point of view, the technical route in the three aspects of perception, decision-making, and execution has been very clear, such as taking the perception layer as an example, everyone waits for the lidar to reduce costs and increase efficiency, and the same is true for the technical route of the line control movement in the execution layer.
The decision-making level with the largest variable has also formed a clear route, and several large mountains are in front of it, which is prohibitive.
The five super manufacturers with chip + system closed-loop capabilities, of which Google's waymo, Tesla and Apple are only for their own use and do not supply outside, the remaining two Intel's mobileye, and Huawei have obvious shortcomings (mobileye is limited above L4, Huawei is too strong as mentioned above), and is abandoned by most manufacturers.
As a result, the new forces and traditional car manufacturers have opened a mix-and-match road. Chips, algorithms, systems, or investment, or shareholding, or self-research, according to their own situation to form their own characteristics.
Players of the self-developed chip route include baiqi and British chip design company imagination in May 2020 to establish a joint venture, and in October 2020, Geely and ARM set up a chip design company, Core Engine Technology. The ideal procurement route and horizon chip + self-developed system, Weilai adopts NVIDIA chip + self-developed system.
The reason why everyone does not tangle or arbitrarily convert methods, the reason is very simple, several super large factories demonstration effect is too strong, Tesla from 2019 began to use their own FSD chips, automatic driving a ride on the dust, the accumulation of data in the sky, super computing power platform, and the cancellation of the extreme behavior of millimeter wave radar, almost became the industry's bottom-up benchmark.
On the other hand, Baidu's Apollo, which began to develop in 2015, will also burn out the barriers with money and time in the top-down platform model, so that every entrant touches the pocket and hesitates.
Any account is an economic account in the end, and now the entrants of the car are all experienced players, feedback to the real news level, in fact, as mentioned above, the main engine factory and tier1 for some technical talents to clear.
There is nothing new under the sun, in fact, the same plot has also happened in Tesla, Tesla's original founder and CTO technology maniac Albert, an engineer, attaches great importance to technology research and development, and the contribution to Tesla is mainly in technology. However, as a CEO who attaches too much importance to technology, he slowly seems to be a little weak in other aspects of corporate management. In Musk's biography "Silicon Valley Iron Man", it was recorded that once, Musk found a financial expert to calculate the production cost of Roadster, and found that Musk once quoted less than $100,000 for the Roadster, and the production cost reached $200,000, more than doubled. In Musk's view, the key at that time should not be entangled in more refined technology, but in cost control and mass production, otherwise the more it sells, the more Musk loses. So Musk didn't hesitate to fire Aberhard.
For the various players of new energy vehicles, the key contradictions at each stage are not the same, obviously, automatic driving is still too advanced for them, and maintaining the current path, to the critical point of technology, the latecomer advantage to catch up, is the choice of most people.
Before dawn, all chicken ribs?
At present, the penetration rate of new energy vehicles in the mainland has reached 19%, and looking back at history, after the penetration rate exceeds 20%, financial leverage will enter the market and accelerate the development of the industry. In the face of the surging market, what is the key contradiction? Take a closer look at the layout of various manufacturers. A very obvious trend is that at this stage, how to better realize electrification and build brand advantages are greater than intelligence.
Tesla is betting on 4680 batteries and building new gigafactories, strategically, or to increase production. On the route, high-nickel ternary batteries have become a mature solution to replace cobalt, from lightweight and endurance to completely crush other brands, Musk in order to straighten out the supply chain, and even promised not to launch new cars in 2022.
The traditional car-making side, after creating the super explosive Wuling Hongguang mini and the euler good cat with monthly sales of 10,000 vehicles, seems to be confirming another idea, the extreme of intelligence can not represent the pain points of all users, build a brand, seize the demand, and create a blockbuster model is king.
On the side of the new forces, in the face of the giants who have entered the market, they do not fully emphasize intelligence, but they are emphasizing the cost performance and product positioning experience, and it seems that Weilai is going to be out of et5 this year.
After experiencing the 2020 radar, algorithm route dispute, 2021 smart car stacking drama, players look back but find that the results are frequent accidents, and scolded on the Weibo hot search, and the other end of the negative feedback cycle: consumers, car critics have long been educated in these years to understand, the so-called free hands-free automatic driving, a short time can not come.
From a financial point of view, automatic driving is a waste, expensive, the direction of low income, the giants rely on the blood transfusion of their main business under a huge cost, and now it is only a small step ahead, the main engine factory to see their own still losing accounts, no one is not clear about this addition, subtraction, multiplication and division. Especially in 2022, when competition is intensifying, Japanese companies are entering the game, Apple is perfecting the system, and everyone imagines that the plot of smart phones will happen again on new energy vehicles. Where is there still time to focus on the so-called autonomous driving in the "future" promising direction?
In the "Medium- and Long-term Development Plan for the Automotive Industry" issued by the Ministry of Industry and Information Technology in 2017, it is clearly stated that by 2025, the assembly rate of new cars in automotive DA, PA, and CA (corresponding to L1, L2, and L3 in the current market) will reach 80%, of which the assembly rate of PA and CA-level new cars will reach 25%, and highly and fully autonomous vehicles will begin to enter the market.
It is precisely because of this statement that the major OEMs and tier1 are taking 2025 as the deadline for the real realization of autonomous driving, but now it seems that the constraints of many parties are very strong.
In order to achieve true automatic driving, communication with the government, road network coordination, and coordination of autonomous driving and non-autonomous vehicles are still in the initial stage because they involve too wide a range.
When greater contradictions appear in front of the forces of all sides, which is more important than the other, everyone has their own scale in their hearts.