Summary of Key Points
This news article focuses on the trend in the intelligent vehicle industry shifting from "single-vehicle intelligence" to "vehicle-cloud integration." Huawei Cloud announced an Intelligent Driving Cloud Solution at the Shanghai Intelligent Vehicle Forum, aiming to address three major challenges faced by automakers: slow data generation, poor model training, and difficulties in mass production and operation. The article also discusses the advantages and challenges of China's intelligent driving industry, as well as cooperation practices between Huawei Cloud and automakers such as Great Wall and Dongfeng Nissan, ultimately pointing towards the transformation direction of "AI-native automakers."
Detailed Analysis
Why Do Intelligent Vehicles Need to Move from "Single-Vehicle Intelligence" to "Vehicle-Cloud Integration?"
Previously, intelligent vehicles relied on "single-vehicle intelligence," meaning the vehicle itself handled computing power and data. However, this approach had significant limitations:
- Insufficient Computing Power: The limited space within a vehicle makes it impossible to install supercomputers, leading to slow response times when dealing with complex road conditions (such as pedestrians, electric vehicles, and traffic lights in cities).
- Limited Data: A single vehicle can collect only a small amount of data, resulting in infrequent model updates (from weekly to daily now).
- Difficult Iteration: OTA (Over-The-Air) updates require substantial data and computing power, making it difficult for vehicles to remain constantly up-to-date.
"Vehicle-cloud integration" essentially provides the vehicle with a "superbrain": it transmits real-time data to the cloud, where massive computing power is used to process and train models, which are then sent back to the vehicle. This not only allows for quick updates of assisted driving functions but also enables multiple vehicles to share data (for example, if Vehicle A encounters new road conditions, Vehicle B can learn from it). Huawei Cloud describes itself as providing the "basic cloud services," enabling automakers to avoid building their own supercomputing centers and focus on developing intelligent driving systems.
Huawei Cloud's New Solution: Three Key Measures to Overcome Challenges
The newly released Intelligent Driving Cloud Solution targets these issues:
- High Data Generation Efficiency: It can process 300,000 pieces of data per day (such as video from roads and sensor data) and retrieve information quickly (e.g., finding a parking lot in seconds), helping automakers obtain useful data more promptly.
- Improved Model Training: The cloud-based infrastructure is tailored for assisted driving, increasing resource utilization by 30% (using less cost to achieve better model accuracy).
- Stable Mass Production and Operation: It uses a "dual-zone" approach for data storage to ensure compliance with regulations and provides a ready-made model toolkit (Shengteng Intelligent Driving SDK) that supports over 60 mainstream assisted driving models and connects with more than 80 hardware partners, accelerating development processes.
For example, after adopting Huawei Cloud's solution, the number of vehicles using the ADS (Assisted Driving System) has surged from 1.7 million to 3 million by the end of the year, with a total assisted driving distance exceeding 10 billion kilometers, and it can now locate parking lots in seconds.
China's Advantages and Challenges in Intelligent Driving
China has clear advantages:
- Diverse Scenarios: Complex urban road conditions generate rich training data.
- Abundant Data: 2 million assisted driving vehicles and 60 million connected vehicles daily contribute to a continuous stream of data for Huawei Cloud.
- Strong Computing Power: Ten thousand Shengteng computing units support model training, with over 30 automakers (such as Changan and Great Wall) collaborating with Huawei Cloud.
However, there are also challenges:
- Mismatch in Technological Progress: Long development cycles for automakers (e.g., 3 years) contrast with rapid AI advancements, leading to potential gaps.
- Inconsistent Standards: Varying technical standards among automakers make cooperation challenging.
- Lagging Regulations: Regulatory frameworks need to keep up with technological developments.
Industry expert Shi Jianhua emphasizes the importance of open collaboration, suggesting that more AI companies should be involved in the automotive sector to establish common standards and drive industry progress.
How Cloud Technology Changes the Driving Experience
Several automakers have shared their success stories with Huawei Cloud:
- Great Wall Motor: Developed a "double-intelligence agent" approach combining on-board and cloud-based intelligence, allowing the vehicle to proactively serve users (e.g., planning routes, reminding of traffic delays, and adjusting the air conditioning temperature).
- Dongfeng Nissan: Collaborated on the "Tianlai·Hongmeng Cockpit," providing intelligent features even in non-electric vehicles (voice control, smartphone integration), with AI improving efficiency across the entire development and sales process.
- Changan Motor: Exploring the integration of vehicles with robots, creating vehicles with sensory capabilities (recognizing emotions) and interactive services (e.g., delivering parcels).
Future Trends: From "Functional Vehicles" to "AI-Native Vehicles"
The future of intelligent vehicles is not about adding AI features but about becoming AI-native companies:
- Extended Value Chain: Automakers will sell not only vehicles but also software services (such as assisted driving subscriptions and smart cockpit upgrades).
- Open Ecosystem: Huawei Cloud promotes open-source collaboration with various partners to create a more intelligent transportation system.
- Ultimate Goals: Safer, more efficient, and more user-friendly transportation experiences.
In summary, the competition in the intelligent vehicle industry is shifting from comparing individual vehicle specifications to evaluating the combined intelligence of vehicles and clouds. Huawei Cloud's solution addresses core challenges for automakers. China's unique scenarios and data advantages, along with an open ecosystem, position the country well to lead in the field of intelligent driving. In the future, your car may understand you better than you do—this is not science fiction but a reality that is already unfolding.