Nvidia Launches Open AI Models to Transform Autonomous Vehicles

Nvidia has introduced a groundbreaking family of open AI models aimed at revolutionizing autonomous vehicles (AV) and robotics during the Consumer Electronics Show (CES) 2026 in Las Vegas. On January 5, company CEO Jensen Huang announced the launch of the Alphamayo family of open-source AI models, which he described as “the world’s first thinking, reasoning, autonomous vehicle AI.”

The Alphamayo initiative is designed to enable AVs to make decisions akin to human reasoning. Huang explained that the goal is to enhance decision-making processes in AVs and to apply similar technologies to physical robots. “Alphamayo is trained end-to-end, literally from camera in to actuation out,” he stated. This means that the system processes sensor input to control steering, braking, and acceleration while also reasoning through potential actions.

Rather than operating directly within vehicles, Nvidia’s Alphamayo will act as large-scale teacher models that developers can customize for their specific AV systems. The core of this new family is the Alphamayo 1, which boasts a sophisticated architecture featuring 10 billion parameters designed for vision, language, and action processing. This model empowers AVs to tackle complex scenarios, such as navigating a traffic light outage or responding to pedestrians, without requiring prior experience.

Huang emphasized the significance of addressing the “long tail of driving,” noting the impracticality of collecting data for every possible scenario that could occur in various countries and situations. Instead, he explained, “These long tails will be decomposed into quite normal circumstances that the car knows how to deal with.”

The underlying code and datasets for Alphamayo are available on Hugging Face, enabling developers to fine-tune Alphamayo 1 into smaller runtime models suitable for vehicle development. Additionally, developers can leverage this technology to create tools that facilitate AV development, including reasoning-based evaluators and automatic labeling systems that tag video data for easier analysis.

Nvidia is also releasing an open dataset containing over 1,700 hours of driving data collected across diverse conditions and geographies. This dataset addresses rare and complex real-world edge cases that AVs may encounter. Alongside this, Nvidia is rolling out AlphaSim, an open-source, end-to-end simulation framework designed for validating AV driving systems. Available on GitHub, AlphaSim offers realistic sensor modeling, configurable traffic dynamics, and scalable closed-loop testing environments.

Looking to the future, Huang expressed confidence in the growth of autonomous vehicles, stating, “In the next 10 years, I’m fairly certain a very, very large percentage of the world’s cars will be autonomous.” He added that the techniques introduced through Alphamayo and simulation can be applied across all forms of robotic systems.

This latest development from Nvidia marks a significant leap forward in the integration of AI into the autonomous vehicle industry, positioning the company as a leader in this evolving field.