Nexar Introduces BADAS AI Model Built on 10 Billion Miles of Real-World Driving Data

Nexar has launched BADAS (Beyond ADAS), a foundation model for vehicle safety and autonomy trained on more than 10 billion miles of real-world driving data, announced in a press release. The model is built using data from Nexar’s global network of over 350,000 connected devices, capturing 60 million annotated driving events.

BADAS is designed to anticipate upcoming road scenarios rather than simply react to visible conditions. According to Nexar, the model outperforms simulation-trained systems in predicting incidents and understanding driver behavior. As new data is collected from connected vehicles, BADAS continuously updates, refining its predictive capabilities.

The model is part of Nexar’s multi-layered technology stack that transforms road video into structured intelligence. This stack includes a crowdsourced sensor network, a proprietary processing pipeline, and predictive AI components that deliver real-time safety insights through APIs and SDKs. Nexar has also developed an interactive driving simulation to demonstrate BADAS’s predictive responses in real-world scenarios.

With BADAS, Nexar positions itself as an AI infrastructure provider for mobility, offering partners tools to train, validate, and deploy safety-critical systems grounded in real-world data.