SHERIDAN, WYOMING – June 17, 2025 – Nvidia has introduced a groundbreaking autonomous vehicle software platform designed to make self-driving technology safer, smarter, and more scalable for automakers, truck manufacturers, robotaxi companies, and mobility startups worldwide.
This new platform, Nvidia Drive, delivers a full-stack solution that combines advanced software with Nvidia’s industry-leading accelerated computing hardware. The result is a robust foundation for AI-powered transportation that is designed to grow alongside evolving technology and regulations.
Smarter autonomy with a full-stack solution
At the heart of Nvidia Drive is a modular, flexible software architecture that supports real-time sensor fusion and continuous over-the-air updates. Whether a manufacturer wants to adopt the complete system or only select features, Nvidia Drive makes it easy to scale and adapt.
Consumers stand to benefit from enhancements in features like:
- Surround perception and automated lane changes
- Active safety interventions
- Smart parking capabilities
These advancements support Level 2++ and Level 3 autonomy today, with a clear path toward even higher levels of automation in the future.
A powerful trio for AI development
To ensure safety and efficiency across the entire development pipeline, Nvidia has built a three-computer ecosystem:
- Nvidia DGX systems and GPUs: For training AI models and developing AV software using vast datasets.
- Nvidia Omniverse and Cosmos on OVX systems: Simulate and generate synthetic driving data, creating real-world scenarios to test and validate AV performance.
- Nvidia Drive AGX: The in-vehicle computer that processes real-time sensor data for safe, high-level autonomous driving.
This integrated approach ensures that AVs are built with a strong foundation and validated through extensive virtual and real-world learning.
Leveraging generative AI for safer decisions
Since most road accidents are caused by human error, Nvidia’s use of AI aims to radically improve safety. Rather than relying on rigid, rule-based systems, Nvidia Drive uses deep learning models trained on massive datasets of human driving behavior.
This means vehicles can now:
- Understand and react with human-like decision-making
- Eliminate the need for predefined rules or modular pipelines
- Continuously learn from both real-world and simulated driving experiences
Thanks to the Nvidia Omniverse Blueprint, developers can convert thousands of human-driven miles into billions of simulated ones, improving vehicle performance while accelerating time to market.
A safety-first mindset with Halos and DriveOS
Earlier this year, Nvidia launched Nvidia Halos, a safety system that spans from cloud infrastructure to in-vehicle applications. It integrates hardware, software, AI models, and safety tools to protect every phase of AV development and deployment.
Complementing Halos is Nvidia DriveOS, a safety-certified operating system (ASIL B/D) designed for autonomous driving. With adaptive sensor support and built-in safety layers, this platform offers peace of mind for consumers and manufacturers alike.
As Nvidia describes it, “the AV stack can deliver the tools, compute power and foundational AI models needed to accelerate safe, intelligent mobility.”
Looking ahead
With Nvidia Drive, the journey toward safer, smarter autonomous vehicles is gaining momentum. By combining powerful AI, scalable software, and an unwavering focus on safety, Nvidia is paving the way for a future where intelligent mobility is not just possible—but dependable.
Learn more at www.nvidia.com.