Nexar's Vera platform has won the 2026 AI TechAward in the Machine Learning Innovation category, awarded by DevNetwork's expert-led Advisory Board. The platform comprises three components: BADAS 2.0, Nexar Apex, and Risk Index. BADAS 2.0 is a collision anticipation model family that achieves 99.4% Average Precision and ranks #1 across all four major industry benchmarks. Nexar Apex is described as the first real-world autonomous vehicle testing standard, providing a shared, reproducible benchmark for autonomous performance against human driving. Risk Index converts 10 billion miles of verified ground truth into predictive roadway risk scores for insurers, fleets, and departments of transportation.
What the platform does
Nexar Vera is built on a network of 350,000 cameras that capture 100 million miles of real driving every month. The company states this produces the world's largest classified archive of naturalistic safety-critical events: 60 million edge-case videos, 10 billion miles of ground truth, and 45 petabytes of verified road intelligence covering 94% of US roads. The platform serves GM, Waymo, Lyft, IBM, NVIDIA, and government infrastructure clients. Nexar emphasizes that it competes with none of them, positioning independence as the core product.
How the award was decided
Award winners were selected from hundreds of nominees by the independent DevNetwork Advisory Board. Selection criteria included technical innovation, notable attention and awareness in the AI ecosystem, and general regard and use by the developer, engineering, and technical communities. Nexar will be presented the award during AI DevSummit 2026, taking place May 27–28 in South San Francisco, CA.
Tradeoffs
Nexar Vera is camera-first and does not use lidar. This is a deliberate architectural choice that reduces hardware cost and complexity, but it also means the system relies entirely on visual data for perception and collision anticipation. The company claims BADAS 2.0 out-scores Waymo's perception stack on edge latency and false-positive rates without lidar, but independent verification of those specific claims is not provided in the announcement. The platform's effectiveness depends on continued access to its camera network and the quality of ground-truth data collected from consumer dash cams.
Bottom line
Nexar Vera's award signals that camera-first, real-world AI is now a benchmark for mass-market autonomy. The platform's combination of collision anticipation, testing standards, and risk scoring provides a verification layer for physical AI systems. For developers and engineers evaluating autonomous driving stacks, Nexar Vera offers an alternative to lidar-dependent approaches, backed by a large real-world data set and independent positioning.