Event Recap | Vehicle-cloud integration : explore more scenarios of AI defined vehicle

From October 21 to 24, the 32nd China Society of Automotive Engineers Annual Conference & Exhibition (SAECCE 2025) successfully concluded in Chongqing. Carlinx CTO ,KevinLIU,delivered the keynote speechtitled “Vehicle-Cloud Integration Expands AI-Defined Automotive Scenarios.”

During his presentation, Liu highlighted that, given the critical transition from.“software-defined” to “AI-defined” vehicles, Carlinx, is leveraging its Hyper Flow MaaS+ data foundation and Hyper Agent scenario-based intelligent platform, to build a “vehicle-cloud integrated + AI domain-intelligent” technology framework. This system provides automakers with flexible configuration, multi-rule sharing, and closed-loop data intelligence solutions, mining the data value across the entire vehicle lifecycle—from R&D, production, supply, sales, to service.

Speech Excerpt:

“Automotive intelligence is shifting from being function-driven to intelligence-driven. This transformation requires a move from the traditional, extensive data acquisition model to a more intelligent and refined approach. Instead of relying solely on massive cloud-based processing, vehicles now need to gain preliminary data-processing edge computing capabilities.

With the rapid advancement of in-vehicle intelligence and computing power, vehicle–cloud interaction is evolving from the exchange of bits to tokens. Centralized processing is no longer efficient or cost-effective, imposing a heavy burden on automakers in terms of network bandwidth, data traffic, and cloud resources.

The new vehicle-cloud integrated data architecture enhances data fluidity while leveraging lightweight distributed computing. This not only makes vehicles smarter but also significantly reduces the automakers’ total cost of ownership (TCO).”

Technological Breakthrough: Four Pillars ofIntelligent Data Foundation

Flexible Configuration & Precise Data Collection

Moving away from the traditional“full upload and cloud cleansing”model that entails high costs, this solution enables the cloud to dynamically and flexibly configure data collection rules on the vehicle side. Hyper Agent can be activated on demand for specific scenarios—such as fault diagnostics or function testing—to capture only the most relevant data. Through edge-side preprocessing and compression, it achieves up to 90% reduction in data traffic and 75% savings in storage costs from the source.

Multi-Rule Sharing & Efficient Reuse

The platform-based design allows data collection and analysis rules created for different business scenarios to be centrally managed and shared in the cloud. For example, diagnostic models developed for “online reliability testing” can share underlying data rules with “AI intelligent diagnosis” applications, eliminating duplicate development efforts. This significantly enhances AI deployment efficiency and iteration speed, achieving collaborative intelligence across business domains.

Data Closed-Loop & Continuous Evolution

By applying congestion-optimization algorithms, transmission latency is reduced by 70%, forming the loop of "collected-analyze-decide-recollect". This empowers vehicle systems with self-learning and evolutionary capabilities. For instance, once the vehicle fault data is collected, the cloud-based diagnostic model can generate a repair plan. Then, the execution result of the repair plan is fed back into model iteration, improving fault diagnosis accuracy up to 92%.

Intelligent Applications & Value Emergence

Given the robust close-loop data foundation, myriads of domain-specific intelligent applications emerge across the entire vehicle lifecycle,transforming raw data into actionable insights. These applications help enterprises unlock greater business value by turning data potential into tangible operational and commercial outcomes.

"We firmly believe that the car of the future will be a continuoisly evovling mobile intelligent agent," concluded Kevin LIU." At Carlinx, we are decdicated to building a powerful digital nervous system for this intelligent entity. Through our vehicle- cloud integrated platform, we aim to work hand in hand with automakers and ecosystem partners to truly achieve our vision of 'making every mile more valuable.'"