VinFast Pivots to Camera-First AI for Scalable Autonomy

The company’s shift centers on the Robo-Car architecture, a departure from the LiDAR-heavy models favored by many competitors. By utilizing seven production-grade cameras and a high-performance computing platform, the system processes 20 trillion operations per second to interpret traffic patterns. This strategy aims to slash the high costs associated with autonomous hardware, making advanced driver assistance accessible to a wider consumer base.

VinFast currently integrates Level 2 features like adaptive cruise control and lane centering into its existing electric vehicle lineup, including the VF 8 and VF 9. However, the roadmap targets a transition to Level 2++ capabilities, which require active human supervision but offer significantly more sophisticated decision-making. Pilot testing for these systems is currently active in controlled zones in Hanoi, with expansion into international markets planned as the software matures through real-world data collection.