Nvidia launches Alpamayo, open AI models that allow autonomous vehicles to ‘think like a human’

At CES 2026, Nvidia introduced Alpamayo, a groundbreaking suite of open-source AI models, simulation tools, and datasets aimed at enhancing the capabilities of autonomous vehicles. This innovative technology is designed to empower these vehicles to navigate complex driving scenarios more like a human, providing a significant leap in the realm of physical AI. Nvidia's CEO, Jensen Huang, heralded the launch as a pivotal moment in AI development, stating, "The ChatGPT moment for physical AI is here—when machines begin to understand, reason, and act in the real world." The centerpiece of this launch is Alpamayo 1, a sophisticated 10-billion-parameter model that enables autonomous vehicles to tackle unusual and intricate situations, such as handling a traffic light failure at a busy intersection, without prior experience. Ali Kani, Nvidia’s vice president of automotive, elaborated on the model’s capabilities, explaining that it breaks down challenges into manageable steps, evaluates every possible outcome, and identifies the safest route. Huang further emphasized that Alpamayo not only processes sensor data to control the vehicle but also articulates the reasoning behind its actions, effectively explaining the rationale for its driving decisions. Developers can access Alpamayo 1’s code on Hugging Face, allowing them to customize it into smaller, more efficient versions for their vehicles. They can also utilize it for training basic driving systems or creating additional tools, such as auto-labeling systems that categorize video data and evaluators that assess the decision-making of vehicles. In conjunction with Alpamayo, Nvidia is releasing an extensive open dataset comprising over 1,700 hours of driving data sourced from diverse environments and conditions. This dataset includes rare and complex real-world scenarios, providing invaluable resources for developers. Additionally, Nvidia is launching AlpaSim, an open-source simulation framework hosted on GitHub, which is designed to replicate real-world driving conditions for comprehensive and safe testing of autonomous systems at scale.