Xiaomi has unveiled the Xring D100, its first self-developed high-performance chip for intelligent driving, marking a major step in the Chinese technology giant’s growing electric vehicle ambitions.
Built using a 3-nanometer process, the D100 is designed specifically for automotive AI workloads. Xiaomi says it is China’s first high-computing-power intelligent-driving chip built on 3nm technology, with commercial deployment planned for 2027.
Reported specifications include a 20-core CPU, a 16-core neural processing unit (NPU) and support for up to 160GB of unified memory. More significantly, Xiaomi says the chip can locally run AI models containing up to 200 billion parameters.
That could give future Xiaomi vehicles substantially greater onboard computing capability, allowing sophisticated AI workloads to be processed inside the vehicle rather than relying on external servers.
The move also gives Xiaomi greater control over its technology stack. Its existing SU7 and YU7 electric vehicles rely on Nvidia hardware for intelligent-driving functions.
Developing its own processor could allow Xiaomi to optimize the chip, software and AI models specifically for its vehicles while potentially reducing costs as production scales.
Xiaomi is not alone. BYD, Nio, Li Auto and XPeng are also developing proprietary automotive processors as Chinese automakers seek greater technological independence. XPeng, for example, has deployed its in-house Turing chips in vehicles and its GX robotaxi platform.
However, one crucial specification remains unknown: Xiaomi has not disclosed the D100’s TOPS performance, making estimates of 700–1,000 TOPS unconfirmed.
Developing custom silicon is also expensive. Xiaomi says it has invested more than 21 billion yuan ($3.1 billion) in its semiconductor programme and employs nearly 3,000 people in its chip division.
If Xiaomi meets its 2027 target, the D100 could become a key component of its strategy to control everything from the AI model to the silicon running it. The future of intelligent vehicles may increasingly depend not just on the car, but on the computer inside it.
