AI-Defined Vehicle Report: How AI Reshapes Vehicle Intelligence?
Chinese OEMs’ AI-Defined Vehicle Strategy Research Report, 2025, released by ResearchInChina, studies, analyzes, and summarizes the concept of AI-defined vehicles, the differences between AI-defined vehicles and software-defined vehicles, the three key elements (data, computing power, and model) of AI-defined vehicles, the strategies and layout of mainstream OEMs in these three elements, how AI enables intelligent vehicle manufacturing, and the AI strategies and layout of mainstream OEMs in areas such as intelligent driving and intelligent cockpit.
AI-defined vehicles refer to a new generation of vehicles that use artificial intelligence (AI) technology as the core driving force to reshape the full lifecycle of vehicles, involving R&D, design, production, usage, and services, in an all-round way. The core of AI-defined vehicles lies in feeding data and training rule-free AI foundation models to improve understanding, perception, and data decision capabilities in complex scenarios. The rapid iteration of AI foundation models marks a turning point from software-defined vehicles to AI-defined vehicles, that is, rule-based intelligent algorithms are being replaced by more flexible core AI technologies. From a technical perspective, "software-defined vehicles" emphasize expanding functionality through software upgrades, while the introduction of AI technology enables vehicle intelligence to break through fixed rules, giving vehicles the ability to learn and grow on their own.
AI-defined vehicles: Advance intelligent vehicles from "usable" to "easy to use": Some functions of software-defined vehicles still remain at the "usable" stage, and the shortcomings in accuracy, stability, and intelligent decision-making significantly affects user experience. AI-defined vehicles will reshape intelligent vehicles in multiple aspects, including intelligent cockpit, intelligent driving, and chassis domains, facilitating the evolution of intelligent vehicle products from functionality to capability. This will help to transform vehicles from a mere transportation mean into a "super agent" or a "smart mobility lifeform".
1. AI-defined Vehicles rely on deep coupling of three key elements: data, computing power, and model.
Data refers to various types of information collected when the vehicle travels and interacts with the external environment. It serves as the "fuel" for AI-defined vehicles, providing the basic materials for algorithm training and optimization. Computing power includes cloud computing centers and vehicle AI chips, which process data and execute computing tasks. It acts as the "engine" of intelligent vehicles, determining the upper limit of system performance. Model refers to a range of computing steps and rules based on AI theory and mathematical models, used to process and analyze data and achieve specific intelligent functions. It serves as the "brain" of vehicles, determining the level of intelligence.
OEMs need to simultaneously deploy all the three elements: In terms of data, they need to establish all-scenario coverage capabilities; in terms of computing power, they need to break the energy efficiency bottleneck of chips; and in terms of model, they need to achieve vehicle-cloud cooperative reasoning. The ultimate form of AI-defined vehicles relies on the deep coupling of the three elements, forming a self-evolving system where "data becomes more refined with use, computing power becomes higher and more efficient, and models improve with training".
2. In rapid iteration of intelligent driving AI, competition over VLA models starts in 2025.
AI technology in intelligent driving evolves and iterates at an exceptionally fast pace, from traditional CNNs to BEV+Transformer (2023), end-to-end (2024), end-to-end+VLM (late 2024), and VLA (2025). VLA marks a paradigm leap in intelligent driving technology from "separation of perception and decision" to "integration of perception, reasoning, and execution".
As an advanced form of traditional end-to-end intelligent driving, VLA (Vision-Language-Action) model addresses three core challenges of current intelligent driving systems through multimodal fusion (vision + language + execution) and chain-of-thought reasoning: global decision capability, breakthroughs in interpretability, and a leap in generalization performance.
Li Auto, Xpeng, Geely, and Xiaomi have all announced plans to gradually introduce VLA in their vehicles starting in 2025. Other OEMs, while adopting different (or similar) technology paths, are not lagging in integrating AI.
2025 may become the "singularity moment" for VLA-based intelligent driving solutions. The adoption of VLA is not just a technological upgrade but a transformation of intelligent vehicles from a mere "tool" into an "agent". In this race, companies with data bases, computing power advantages, and popular vehicle models will have a say in the automotive industry in the next decade. For consumers, more humanized mobility experience and fiercer market competition will be dual background colors in China's intelligent vehicle industry in 2025.
3. OEMs are quickening their pace of deploying AI and applying AI in vehicles.
Seen from Li Auto’s layout in AI-defined vehicles, since 2024, the company has entered a boom period of vehicle intelligence. It has rolled out industry’s first end-to-end + VLM dual-system intelligent driving, and "parking space to parking space" intelligent driving, and plans to mass-produce and implement its next-generation autonomous driving architecture, Mind VLA, in Q3 2025.
Li Auto initiated its vehicle operating system R&D project in 2021. It input a 200-person team and over 1 billion yuan in R&D expense, and has completed solution selection, architecture design and implementation. The first version was mass-produced and used in vehicles in 2024. At the 2025 ZGC Forum Annual Conference in March 2025, Li Xiang, Chairman of Li Auto, announced that the company would open-source its vehicle OS. By Li Auto’s estimates, the open-source Halo OS could save the automotive industry 10-20 billion yuan annually by eliminating redundant R&D investments, further accelerating the development of AI-defined vehicles in China.
Since the beginning of 2025, Geely has fully embraced AI, positioning itself as a popularizer of intelligent vehicle AI technology. At CES 2025, Geely unveiled its “Full-Domain AI for Smart Vehicles” technology system. The company believes that true intelligent driving is not just about stacking features but AI enablement.
In the run-up to its product launch in March 2025, Geely partnered with Lifan Technology to establish a joint venture, Chongqing Qianli Intelligent Driving Technology Co., Ltd. Yin Qi, Chairman of Qianli Technology, is also a co-founder of Megvii, one of China’s "Four AI Dragons".
According to Yin Qi, AI technology is transitioning from L2 "reasoner" to L3 "agent", and it is the widespread belief in the industry that 2025 is the year of AI application explosion. This trend will first ignite "AI + vehicle".
How will AI define vehicles? Clues may be found in cooperation between Geely and Qianli Technology in three key areas: Ultra-Natural User Interface (NUl), Autonomous Driving & Execution (ADE), and Scaling Law for Al on EV.
New Energy Vehicle 800-1000V High-Voltage Architecture and Supply Chain Research Report, 2025
Research on 800-1000V Architecture: to be installed in over 7 million vehicles in 2030, marking the arrival of the era of full-domain high voltage and megawatt supercharging.
In 2025, the 800-1000V h...
Foreign Tier 1 ADAS Suppliers Industry Research Report 2025
Research on Overseas Tier 1 ADAS Suppliers: Three Paths for Foreign Enterprises to Transfer to NOA
Foreign Tier 1 ADAS suppliers are obviously lagging behind in the field of NOA.
In 2024, Aptiv (2.6...
VLA Large Model Applications in Automotive and Robotics Research Report, 2025
ResearchInChina releases "VLA Large Model Applications in Automotive and Robotics Research Report, 2025": The report summarizes and analyzes the technical origin, development stages, application cases...
OEMs’ Next-generation In-vehicle Infotainment (IVI) System Trends Report, 2025
ResearchInChina releases the "OEMs’ Next-generation In-vehicle Infotainment (IVI) System Trends Report, 2025", which sorts out iterative development context of mainstream automakers in terms of infota...
Autonomous Driving SoC Research Report, 2025
High-level intelligent driving penetration continues to increase, with large-scale upgrading of intelligent driving SoC in 2025
In 2024, the total sales volume of domestic passenger cars in China was...
China Passenger Car HUD Industry Report, 2024
ResearchInChina released the "China Passenger Car HUD Industry Report, 2025", which sorts out the HUD installation situation, the dynamics of upstream, midstream and downstream manufacturers in the HU...
ADAS and Autonomous Driving Tier 1 Suppliers Research Report, 2025 – Chinese Companies
ADAS and Autonomous Driving Tier 1 Suppliers Research Report, 2025 – Chinese Companies
Research on Domestic ADAS Tier 1 Suppliers: Seven Development Trends in the Era of Assisted Driving 2.0
In the ...
Automotive ADAS Camera Report, 2025
①In terms of the amount of installed data, installations of side-view cameras maintain a growth rate of over 90%From January to May 2025, ADAS cameras (statistical scope: front-view, side-view, surrou...
Body (Zone) Domain Controller and Chip Industry Research Report,2025
Body (Zone) Domain Research: ZCU Installation Exceeds 2 Million Units, Evolving Towards a "Plug-and-Play" Modular Platform
The body (zone) domain covers BCM (Body Control Module), BDC (Body Dom...
Automotive Cockpit Domain Controller Research Report, 2025
Cockpit domain controller research: three cockpit domain controller architectures for AI Three layout solutions for cockpit domain controllers for deep AI empowerment
As intelligent cockpit tran...
China Passenger Car Electronic Control Suspension Industry Research Report, 2025
Electronic control suspension research: air springs evolve from single chamber to dual chambers, CDC evolves from single valve to dual valves
ResearchInChina released "China Passenger Car Elect...
Automotive XR Industry Report, 2025
Automotive XR industry research: automotive XR application is still in its infancy, and some OEMs have already made forward-looking layout
The Automotive XR Industry Report, 2025, re...
Intelligent Driving Simulation and World Model Research Report, 2025
1. The world model brings innovation to intelligent driving simulation
In the advancement towards L3 and higher-level autonomous driving, the development of end-to-end technology has raised higher re...
Autonomous Driving Map (HD/LD/SD MAP, Online Reconstruction, Real-time Generative Map) Industry Report 2025
Research on Autonomous Driving Maps: Evolve from Recording the Past to Previewing the Future with "Real-time Generative Maps"
"Mapless NOA" has become the mainstream solution for autonomous driving s...
End-to-End Autonomous Driving Research Report, 2025
End-to-End Autonomous Driving Research: E2E Evolution towards the VLA Paradigm via Synergy of Reinforcement Learning and World Models??The essence of end-to-end autonomous driving lies in mimicking dr...
Research Report on OEMs and Tier1s’ Intelligent Cockpit Platforms (Hardware & Software) and Supply Chain Construction Strategies, 2025
Research on intelligent cockpit platforms: in the first year of mass production of L3 AI cockpits, the supply chain accelerates deployment of new products
An intelligent cockpit platform primarily r...
Automotive EMS and ECU Industry Report, 2025
Research on automotive EMS: Analysis on the incremental logic of more than 40 types of automotive ECUs and EMS market segments
In this report, we divide automotive ECUs into five major categories (in...
Automotive Intelligent Cockpit SoC Research Report, 2025
Cockpit SoC research: The localization rate exceeds 10%, and AI-oriented cockpit SoC will become the mainstream in the next 2-3 years
In the Chinese automotive intelligent cockpit SoC market, althoug...