Who would have thought that by 2024, the intelligent driving industry would usher in new entrepreneurs?
Leifeng.com's "New Intelligent Driving" learned that an emerging high-end intelligent driving solution company recently landed in the Shunyi base of the Beijing-Hebei Demonstration Zone of National Intelligent Vehicles and Intelligent Transportation, the company is called IFAUTO Yifu Zhixing, and the founder is Li Xiang, the former director of Cambrian Xingge Autonomous Driving.
At this time, IFAUTO aims at L3 high-end intelligent driving and brings a new intelligent driving solution.
Li Xiang introduced to Leifeng.com's "New Intelligent Driving" that IFAUTO will use pure visual technical routes and end-to-end models to launch a mapless intelligent driving solution covering all scenarios (including high-speed, urban and parking scenarios) with a single solution, which can be opened nationwide once it is mass-produced.
This solution will be very cost-competitive. Li Xiang said that in order to control the price of the full set of solutions below 5,000 yuan, the new company's intelligent driving program adopts an original form of software and hardware, which will be different from the existing intelligent driving program in the industry. After the new company's intelligent driving program matures, the price is expected to hit the 3,000 yuan level.
At the level of no diagram, the difference between IFAUTO and the existing intelligent driving solution lies in two aspects.
First, at this stage, although the industry has been "removing high-precision maps" for a long time, from the actual landing situation, most players are only "light maps" or use customized maps, while IFAUTO Yifu Zhixing "real no map" solution only needs traditional navigation information, and can use perception and other technologies to restore the scene during driving. Li Xiang said that the new company's intelligent driving program will completely get rid of the limitations brought by high-precision map qualifications, update speed, and cost.
Second, Tesla's no-map relies on the form of "crowdsourced maps", that is, a large number of mass-produced vehicles act as map collection vehicles, and obtain their own "high-precision maps" by superimposing these collected information.
"However, Tesla's method has high requirements for data consistency and infrastructure construction of car companies, and is not suitable for domestic car companies. Li Xiang said that IFAUTO's solution will be more in line with the needs of the industry and car companies.
Li Xiang himself has rich experience in intelligent driving. Since 2017, Li Xiang has joined the intelligent driving industry as the technical leader, and has successively held important positions in intelligent driving software and hardware companies such as Momenta, SenseTime, Cambrian Xingge, etc., in addition to being in charge of research and development, he also participates in engineering, product, and business direction, and has a relatively mature perspective on technology and business.
Xingge is an intelligent driving chip subsidiary established by Cambrian in January 2021, and its shareholders include many major companies in the industry, such as SAIC, NIO, CATL, and Bosch Ventures. Cambrian Xingge has two main products, one is the SD5223 chip for the L2+ market, and the other is the SD5226 series for the L4 market, which can support vehicle-end training. At the end of 2022, Cambrian and Xingge were added to the Entity List by the United States, and entering 2023, Cambrian Xingge has experienced a lot of turmoil.
As one of the founding members of Cambrian Xingge, Li Xiang completed the development and implementation of the intelligent driving solution for urban and high-speed scenarios within two years of the establishment of Cambrian Xingge. At the same time, he led the team to go deep into the early design and later verification of on-board chips, and created a domain controller and corresponding solutions based on Cambrian SD5223.
Prior to joining Cambrian, Li Xiang served as the deputy director of autonomous driving technology at SenseTime. During his time at SenseTime, Li Xiang spent more than a year rebuilding the software and hardware architecture of SenseTime's autonomous driving, and completed the development of the Robotaxi project, providing technical support for the subsequent incubation of SenseTime's autonomous driving minibuses, and the related projects and technology accumulation were named "Jueying" when SenseTime went public in 21 years.
Li Xiang joined Momenta in 2017 as the head of the L4 parking department. The project he is responsible for made its debut in 2018, and in 2019, Momenta signed a mass production order with Geely for parking-related technology, and finally installed it on the Geely Xingyue L model, which is also the first batch of mass production of intelligent driving in China.
Before officially entering the autonomous driving industry, Li Xiang also worked in technology development at Zero Zero Infinity. Zero Zero Unlimited is a drone company that directly targets C-end users.
In recent years, the intelligent driving industry has mainly focused on MPI to take over the mileage to improve safety, but there is still a lot of room for improvement in comfort and intelligence directly related to the C-end user experience.
Thanks to its experience in making C-end products at ZeroZero, IFAUTO's solution will not only focus on cost performance, but also optimize from the aspects of user interaction and ease of use.
He revealed to the new intelligent driving that the core team of IFAUTO has been formed, and the members are from Momenta, SenseTime, SAIC, Xiaopeng and other well-known companies in the industry, with rich experience in mass production R&D and management, and have the ability and confidence to complete the mass production of the new company's intelligent driving solution.
At present, IFAUTO has reached a preliminary cooperation intention with a local domestic car company and an overseas car company, and in terms of ecological partners, the new company has established cooperative relations with domestic chip companies and domestic technical service providers, which will be announced later.
Compared with the L2+ main focus of most car companies and suppliers, IFAUTO has gone one step further and has to do L3 high-end intelligent driving.
This also means that IFAUTO not only completes the L3 in the minds of consumers, that is, the experience is better than L2 and L2+ products in terms of safety, comfort and intelligence, but also completes the L3 in the technical definition, that is, under applicable conditions, allows users to hand over the driving right to the vehicle to make independent decisions and operates, so as to achieve "hands-off and eyes-free".
In November 2023, the four ministries and commissions issued the "Notice on Carrying out the Pilot Work on the Access and Road Access of Intelligent Connected Vehicles", officially putting forward specific requirements for the access specifications of L3/L4 autonomous driving and carrying out access pilots. At present, test licenses have been issued to many car companies, including BYD, Changan, AVATR and Zhiji, to promote L3 on the road.
Li Xiang said that IFAUTO has been done according to the L3 standard in the research and development stage.
In the process of forming a new company, Li Xiang often receives such questions: the wave of autonomous driving entrepreneurship began in 2016, and now the industry has become rational and the market pattern has emerged, is it too late to start a business this year?
Li Xiang believes that L3 intelligent driving has only just reached the inflection point, and the real flashpoint has not yet arrived.
From a business point of view, urban NOA has just ushered in the tide of getting on the car, and C-end users have just begun to get in touch with high-end intelligent driving functions, from functional experience to high-end intelligent driving to realize real charge commercialization, and at least 3-5 years of user training period will be experienced in the middle.
Taking the annual data at the end of 2023 as an example, the penetration rate of L2 intelligent driving (including L2+ and L2++) in the new energy market is 50%, while the penetration rate of L3 in the new energy market is only about 5%.
From a technical point of view, the implementation of technology requires an experimental period. Whether you can use high-precision maps or lidar is a dual consideration of technology and business. Over the years, companies engaged in the research and development of high-end intelligent driving have taken a lot of detours, and most of the solutions on the market belong to the process of industry development, whether in terms of sensor configuration, algorithm design, data closed-loop process, etc., there is still a big gap with the end of the industry.
Li Xiang said that the choice of entrepreneurship at this stage is also based on the consideration of overall technical risk, landing clarity and profit cycle, and the entry into the game at this time is on the best curve of technology profitability.
In his opinion, the current development of the entire high-end intelligent driving solution market is uneven, the head self-developed manufacturers have landed in major cities without a map of the urban NOA, the middle and tail manufacturers are actively catching up, in the new models have launched lidar optional configurations, no manufacturer is willing to miss the high-end intelligent driving ticket, so there are still many opportunities for the new company's intelligent driving program.
There is no doubt that despite the current cold investment and financing in the primary market, autonomous driving is experiencing unprecedented certainty - while technology and industry are maturing, intelligent driving is becoming more and more important in users' car purchase decisions, and more and more car companies are showing enthusiasm for intelligent driving.
On the other hand, Tesla's FSD entry into China is imminent. On the morning of April 28, Tesla CEO Elon Musk arrived in Beijing to discuss the entry of FSD into China. And on April 20, Musk said on the social media platform X that Tesla may soon offer the full self-driving software feature FSD to Chinese customers. The addition of Tesla's FSD will further accelerate the popularization of high-end intelligent driving in China and drive the development of the entire autonomous driving industry.
After more than a decade of development, the autonomous driving industry is still thriving, and new variables are still emerging.