Domestic suppliers changed from supporting role to leading role at Shanghai Auto Show, and the end-to-end intelligent driving route was further deepened. World model and VLA technology became the key trends, and L3 became the core topic of the industry.
Although the craze of "all-people intelligent driving" has been cooled down by supervision, many companies have adjusted their expression to "assisted driving", but Chinese and foreign intelligent driving companies and some OEMs are still actively promoting the landing preparation of L3 or even L4.
Strictly speaking, L3 is regarded as a stress test of one’s own ability, which involves the competition of discourse power. China local intelligent driving company is facing a technological breakthrough, and L3 is imminent. Twenty-three supply chain enterprises, including Huawei, Horizon and Momenta, collectively entered the vehicle pavilion, and the exhibition area of automotive technology and supply chain increased to 100,000 square meters, accounting for nearly one third.
Zhijia has formed a global industrial chain with local startups in China as the core. International giants such as Bosch, Continental and Denso signed contracts with Horizon to build a smart driving scheme based on local chips. With the cooperation with Black Sesame Intelligence and Face-wall Intelligence, Intel laid out the ecology of China automobile industry.
Japanese cars represented by Toyota and Nissan, and brands such as GM and Audi fully embrace China Zhijia Company. Momenta reached cooperation with six major brands, and the "Intel Inside" model in the PC era reappeared in the field of smart cars. OEMs gradually accept the fact that suppliers improve their intelligent driving ability, and the traditional Tier 1 model has changed.
Most OEMs give up self-research and turn to external cooperation. On the technical route, "one-stage end-to-end" evolved into VLA architecture, and the discussion focused on the choice of expert system or generalist system and the technological leap from L3 to L4.
Momenta shows a large model of R6 flywheel based on reinforcement learning, which is planned to be launched in the second half of the year. Its assisted driving scheme covers the global market and supports cartographic technology. Robotaxi business is expected to enter the trial operation stage at the end of the year.
Xiaoma Zhixing released the seventh-generation Robotaxi system, which achieved 100% popularity of car-level hardware and reduced the cost by 70%. The system is suitable for many models and has all-weather operation ability. Peng Jun said that 2025 is the first year of mass production of Robotaxi.
Zhuoyu Technology has launched a new type of blind-supplementing LiDAR "knowing the week", which strengthens the combination of learning and world model to improve safety. Shang Tang Jueying shows the R-UniAD scheme of generative intelligent driving, emphasizing the synergy between reinforcement learning and world model.
Guangzhou Automobile Group cooperated with Didi to launch Robotaxi for pre-installation mass production, and planned to deliver it in mass production by the end of 2025. Huawei Gankun Zhizhi drives the ADS 4 high-end version to position the high-speed commercial L3 scheme, and canoe Zhihang promises to share the responsibility with the OEM.
Traditional Tier 1, such as Bosch, is deeply involved in the China market, providing full-stack solutions covering technologies such as urban assisted driving, brake-by-wire and steering. An Bofu builds the localization full-stack capability, and the software and hardware decoupling design supports flexible customization.
The Shanghai Auto Show also showcased emerging fields such as humanoid robots and flying cars. Citigroup predicts that the market size of humanoid robots will reach 7 trillion US dollars in 2050. AI technology promotes the transformation of automobile industry to multi-scene intelligent terminal.
Disclaimer: The content of this article is automatically generated by an open intelligent model for reference only.