VinFast Bets on Camera-Only Self-Driving in Partnership with AI Firm Autobrains

In a move that could reshape the market for affordable self-driving cars, Vietnamese automaker VinFast has entered a strategic partnership with Israeli AI company Autobrains. Announced on January 26, 2026, the collaboration focuses on developing a low-cost autonomous system that relies solely on cameras, a direct challenge to the approach used by Tesla.

The core of the project is a "Robo-Car" architecture designed to achieve advanced autonomous capabilities without expensive sensors like LiDAR. Instead, it uses seven standard cameras and a dedicated processing chip. Pilot programs are already running in Hanoi, using modified VinFast VF 8 and VF 9 SUVs.

VinFast's current vehicles offer Level 2 driver assistance. The new system, classified as L2++, is a significant step toward hands-free operation. The ultimate goal is scalable Level 4 autonomy, where the car handles all driving in certain conditions, at a price point aimed at mainstream buyers.

"This partnership is about making advanced driving technology accessible," said Nguyen Van Duong, VinFast's Global Deputy CEO for ADAS/AD Research. "We are moving it from early adopters to everyday drivers."

For VinFast, which has faced sales and technical challenges in the U.S. market, the partnership offers a potential path to differentiation and cost reduction. Autobrains brings its "Liquid AI" technology, a self-learning system backed by over 250 patents. The company has financial backing from VinFast, Toyota Ventures, and BMW i Ventures, among others.

The initiative represents a clear strategic pivot. While VinFast has other partnerships for sensor-heavy robotaxis, the Autobrains deal prioritizes a minimalist, camera-first vision for personal vehicles. If successful, it could position VinFast as a leader in bringing capable self-driving systems to emerging economies, where the company is building new factories. The next phase will involve expanding pilot tests and integrating the technology into production models.

Source: Webpronews