Nexar Unveils BADAS: Breakthrough AI Model Sets New Safety Benchmark for AV and Driver Assistance Systems

Trained on 10+ billion miles of real-world driving data, BADAS proves the future of mobility intelligence depends on what happens on actual roads—not in simulation.

NEW YORK, Oct. 20, 2025 /PRNewswire/ -- Nexar, a leader in AI-powered mobility solutions, today announced the launch of BADAS (Beyond ADAS), a foundation model that redefines how the industry approaches vehicle safety and autonomy. Built on the largest open dataset of real‑world driving events ever assembled, BADAS delivers measurable gains in incident prediction accuracy and behavioral foresight—outperforming traditional, simulation‑trained models in critical safety benchmarks.

Unlike models developed in labs or virtual environments, BADAS is trained on Nexar's unmatched real‑world vision network—spanning over 10 billion driven miles and 60 million annotated edge-case events captured through Nexar's proprietary global sensor network of 350,000+ connected devices. It is designed to anticipate what's next on the road, not just react to what's in view.

"This is a defining moment for Nexar and the future of mobility," said Zach Greenberger, CEO of Nexar. "BADAS isn't just a model—it's a statement. It proves that the road to safety and autonomy runs through the real world. By building intelligence on top of our unmatched global data network, Nexar is now moving up the stack—from raw road data to deployable, safety-critical AI infrastructure."

The Challenge the Industry Has Yet to Solve

For more than two decades, automakers and tech companies have promised that software and AI would reduce crashes and save lives. Yet road safety has not improved in the way we expected. The critical gap? Traditional approaches lack exposure to the full spectrum of real-world driving behavior.

While AI has transformed digital life—powering chatbots, recommendation systems, and automation—it has done little to protect the physical one. Pedestrian deaths continue to rise, driver‑assist systems remain reactive, and autonomous vehicles are still confined to limited, geofenced zones.

However, if there is one company that has built physical AI with massive impact, it is Tesla: they have proved that AI powered by massive volumes of real-world data can drive significant progress in safety and autonomy on broad operational design domains. Nexar is now bringing that same advantage to the rest of the industry—enabling every OEM, insurer, and fleet operator to train, validate, and deploy systems grounded in real‑world experience.

Solving safety requires models exposed to the full diversity of human driving—the long tail of rare, unpredictable, and behaviorally rich events that simulations cannot replicate. This is exactly the kind of data that BADAS was trained on.

A New Approach: Nexar BADAS

That's why today, Nexar is introducing BADAS (Beyond ADAS), the first foundation model designed to continuously learn from real‑world driving at global scale. As Nexar's network of connected vehicles captures new scenarios on the road, BADAS adapts—constantly improving its understanding of how people, environments, and infrastructure interact.

The model has already demonstrated state‑of‑the‑art performance in collision prediction and driver behavior modeling, establishing real‑world data as the new standard for building and validating mobility AI.

In short: the road to safe autonomy doesn't start in a lab—it starts on the road.

From Data to Deployment: The Nexar Stack Behind BADAS

BADAS is the pinnacle of the Nexar Stack, a multi‑layered system that transforms real‑world driving signals into deployable AI:

• Crowdsourced Sensor Network – Hundreds of thousands of connected vehicles form Nexar's distributed "eyes on the road," generating continuous visibility into traffic, infrastructure, and safety events worldwide.
• Real‑World Data Engine – Nexar's proprietary annotation and processing pipeline transforms raw road video into structured intelligence used to train and validate foundation models.
• Predictive AI and Foundational Models – BADAS anticipates risk and behavior seconds before it occurs—enabling earlier interventions and safer systems across fleets, insurers, and cities.
• Real‑Time Safety Solutions – Through APIs and SDKs, Nexar delivers live, ground-truth road intelligence—crash alerts, work‑zone detection, and risk assessments—to mobility partners everywhere.

Together, these layers enable Nexar to move up the stack—turning real‑world road intelligence into operational AI that serves both industry partners and the public good.

Experience BADAS in Action

To demonstrate BADAS's predictive capability, Nexar created an interactive driving simulation that lets users experience how the model anticipates and responds to real‑world scenarios in real time—highlighting the difference between human drivers and AI trained on reality rather than simulation.

Access the experience here: https://www.nexar-ai.com/badas

"For years, mobility AI has relied on approximations—limited data, closed environments, and controlled assumptions," said Eran Shir, Nexar co-founder and Chief Product Officer. "BADAS changes that. It's built on the unpredictability of real streets, with the richness and diversity of behavior that simulation simply can't replicate. This is what it takes to move from theoretical safety to real-world impact."

With BADAS, Nexar is turning years of real-world driving into predictive, deployable intelligence. This marks a new chapter: Nexar as a foundational AI infrastructure provider for mobility.

From powering proactive APIs to accelerating infrastructure response and validating autonomy stacks, Nexar's platform translates the complexity of real roads into scalable, safety-critical AI. BADAS is the clearest example yet: a safety-critical model built entirely on real-world context—now available to partners ready to build, validate, and scale the next generation of ADAS and AV systems.

For more information visit https://www.nexar-ai.com/

About Nexar

Nexar turns cars into vision sensors to understand the world. Its platform powers vision-connected services and apps at scale, making new vision-based applications for better driving powered by a crowd-sourced vision feed. Using anonymous, aggregated data captured from this network, Nexar has developed a portfolio of vision-based data services for the autonomous vehicle industry and public and private sector partners to make roadways safer and more efficient. More information at https://www.nexar-ai.com/.

Logo - https://mma.prnewswire.com/media/2799479/Nexar_Logo.jpg

SOURCE Nexar