NVIDIA Drive AV software will debut in the all-new Mercedes-Benz CLA, bringing enhanced Level 2 point-to-point driver assistance to U.S. roads by the end of 2026. The CLA is the first production vehicle to feature NVIDIA’s complete AV software stack on Mercedes’ new MB.OS platform, marking a significant step in the automaker’s AI-driven mobility strategy.
What’s New in the Mercedes-Benz CLA
The 2026 Mercedes-Benz CLA integrates NVIDIA’s full-stack DRIVE AV software, powered by NVIDIA DRIVE AGX accelerated compute and supported by AI infrastructure. Mercedes refers to the system as “Level 2++” to distinguish its expanded capabilities from standard Level 2 systems.
Key features of MB.DRIVE ASSIST PRO include point-to-point urban navigation through complex city environments, advanced proactive collision avoidance, automated parking in tight spaces, and cooperative steering between driver and system. The vehicle uses 10 cameras and five radar sensors to process perception and decision-making in real time.
The CLA recently earned a five-star EuroNCAP safety rating, with MB.DRIVE active safety features contributing to accident mitigation scores. Over-the-air updates will enable future feature enhancements, available through the factory or Mercedes-Benz store.
How NVIDIA’s Dual-Stack Architecture Works
NVIDIA DRIVE AV employs a dual-stack design: an AI end-to-end stack for core driving plus a parallel classical safety stack built on NVIDIA Halos. This redundancy ensures fail-safe checks and keeps the vehicle within defined safety parameters.
The AI models interpret traffic holistically, understanding vulnerable road users like pedestrians and cyclists. The system can navigate lane selection, execute turns, and follow routes in unfamiliar areas while responding proactively to prevent collisions.
Ali Kani, NVIDIA’s vice president of automotive, stated the technology makes “every vehicle programmable, updatable and perpetually improving through data and software”. The CLA demonstrates what NVIDIA calls an “AI-defined vehicle” trained on synthetically generated data and neural reconstruction.
Why This Matters for the Auto Industry
NVIDIA’s production launch with Mercedes-Benz signals intensifying competition in AI-powered driver assistance, particularly against Tesla’s Full Self-Driving system. Unlike Tesla’s camera-only approach, NVIDIA combines cameras and radar for sensor redundancy.
The partnership extends beyond software. NVIDIA and Mercedes-Benz are transforming vehicle manufacturing using NVIDIA Omniverse digital twins to design and optimize assembly lines virtually. This “digital-first” approach reduces downtime and accelerates factory iteration.
NVIDIA’s three-computer architecture connects cloud training (NVIDIA DGX systems), simulation validation (Omniverse and Cosmos platforms), and in-vehicle compute (DRIVE AGX). Developers transform real-world data into billions of simulated miles before deployment.
What’s Next for NVIDIA and Mercedes
What is NVIDIA Drive AV software?
NVIDIA Drive AV is a full-stack autonomous vehicle software that uses AI end-to-end models for driving alongside a classical safety stack with NVIDIA Halos. It enables Level 2++ driver assistance with point-to-point urban navigation and proactive collision avoidance.
When does the Mercedes-Benz CLA with NVIDIA software launch?
The Mercedes-Benz CLA with NVIDIA Drive AV software is expected to launch in the U.S. by the end of 2026. It’s the first production vehicle on Mercedes’ MB.OS platform to feature the complete NVIDIA AV stack.
What is Level 2++ autonomous driving?
Level 2++ (or “Level 2 optimized”) refers to enhanced Level 2 systems with expanded capabilities beyond standard driver assistance. It includes point-to-point navigation on highways and city streets while keeping the driver alert and in control.
How many sensors does the Mercedes CLA autonomous system use?
The Mercedes-Benz CLA uses 10 cameras and five radar sensors for perception and decision-making. The NVIDIA DRIVE AGX compute processes sensor fusion in real time to handle complex urban and highway scenarios simultaneously.
