How Do NVIDIA Chips Power Alibaba’s AI Car Systems?

At the Beijing Auto Show, Alibaba said its Qwen artificial intelligence model will run in vehicles built by brands such as BYD and a Volkswagen collaboration. The company said, “The model will run on Nvidia’s automotive chip system and is designed to function even with limited network connectivity.”

That chip system comes from NVIDIA. Its DRIVE Orin platform runs real time perception and planning in the car. In Guangzhou, the chip already powers fully driverless robotaxis through work with WeRide. The same hardware now supports in car assistants that respond to voice, process data locally and connect to cloud services when needed.

Alibaba explained how the system works. “The system combines on-device processing with cloud-based computing to interpret voice inputs, plan multi-step tasks and connect to services such as payments and navigation.” The car processes speech on the chip, then pulls extra data from the cloud when available.

This setup matters because network coverage can drop in tunnels or busy streets. Running core tasks on the NVIDIA chip keeps the system responsive when the signal goes down.
 

How Does The In Car AI Work For Drivers?

 
Alibaba said the goal is to turn the car into a service hub. “Select models will allow drivers to order food delivery, book hotels, buy attraction tickets and track packages through voice commands.”

That shows how far in car systems have come. The assistant goes past navigation or music and handles everyday tasks that usually sit on a phone. It connects digital services in one system.

To make this work, the AI needs to understand context. A request like booking a hotel involves location and timing. The Qwen model breaks that into steps, checks available services and completes the task through connected platforms.

The NVIDIA chip processes the voice input and runs the AI model’s core functions. Cloud systems deal with larger data requests such as searching listings or confirming bookings. This split keeps response times fast in the car.

Automakers are adding these features as competition in the electric vehicle market grows. Software now helps attract buyers, especially in China where brands compete on digital features as much as hardware.
 

Why Is NVIDIA’s Guangzhou Work Important Here?

 
NVIDIA’s work in Guangzhou shows how this technology reaches real roads. The company has built a local network covering hardware and engineering teams. Its DRIVE Orin chip already runs in production vehicles and robotaxis in the city.

WeRide operates a 24 hour fully driverless robotaxi service in Guangzhou using DRIVE Orin. That service runs without safety drivers in designated areas, showing how the chip deals with complex traffic conditions in real time.

NVIDIA has also backed WeRide with an investment valued at over $24 million as of early 2025. That connection gives access to real world driving data and testing environments.
 

 
Local infrastructure is also important here. Giga Computing has rolled out NVIDIA RTX PRO servers in Guangzhou to support AI training and simulation. These servers train models that understand traffic scenes before they are used in vehicles.

This creates a full cycle. Data from cars goes into training systems, improves the models and returns to vehicles through updates.
 

Which Car Brands Are Getting This System?

 
Alibaba listed a group of automakers planning to use Qwen in their vehicles. These include Geely, Li Auto, Changan Automobile, Dongfeng Motor, BAIC Group, Great Wall Motor and SAIC Motor.

SAIC’s work with Volkswagen means the system will also appear in shared models. Earlier in the year, FAW Group’s Hongqi brand added Qwen to its Hongqi HS6 plug in hybrid.

Each brand will adjust the system to its own interface, but the core technology stays the same. NVIDIA provides the computing base and Alibaba supplies the AI model and service connections.

Other companies are working on related systems. A China specific version of Audi said its E7X electric SUV will use AI features from ByteDance’s Doubao and iFlyTek. Cadillac has also shown a model with voice assistant features connected to Doubao.

It is not yet known if these AI systems will appear in cars sold outside China.
 

What Does This Mean For In Car Tech?

 
This setup shows a new way cars are built. Hardware such as chips works together with software models and online services. The result is a car that acts more like a connected device.

NVIDIA runs the heavy computing tasks in the vehicle. Alibaba’s Qwen model adds the intelligence that understands requests and connects them to services.

They create a system together, that can manage driving tasks and daily activities in one place. The car becomes a platform for digital services as much as transport.

With more brands adopting these systems, competition will focus on how well these assistants work in real life. Speed, accuracy and reliability in poor network conditions will matter more than feature lists.