Nvidia has open sourced Alpamayo R1, a full-scale autonomous driving model on GitHub and Hugging Face.
Nvidia has released “the world’s first industry-scale” open-source software for self-driving development, launching the autonomous driving model Alpamayo R1 on GitHub and Hugging Face. The move marks a significant shift in an industry dominated by proprietary stacks and positions Nvidia as a champion of transparent, community-driven autonomy research.
Built on a vision-language-action (VLA) architecture, Alpamayo R1 enables vehicles to detect obstacles, reason through complex road scenarios, and respond safely. Nvidia noted that “reasoning gives autonomous vehicles the common sense to drive more like humans do.” The model supports behaviours such as avoiding bike lanes in pedestrian-heavy zones and aims to accelerate progress toward Level 4 autonomy, where vehicles can drive independently without human oversight.
Researchers can customise Alpamayo R1 for non-commercial uses, strengthening global access to high-end autonomous driving technology. The release expands Nvidia’s Nemotron family of open technologies, further opening safety-critical AI research to academia and independent developers.
The announcement lifted Nvidia shares on Monday, underscoring investor confidence in its deepening role in autonomous mobility. The company continues to expand its ecosystem through partnerships with Lucid, which is integrating Nvidia’s DRIVE AGX and DriveOS, as well as WeRide, backed by Nvidia, and Uber, whose real-world driving data is helping refine Nvidia’s Cosmos environment-understanding models.
Named after a mountain in Peru, Alpamayo R1 emerges as a technical and strategic signal: Nvidia is democratising access to autonomous vehicle intelligence while challenging the dominance of closed self-driving systems. Wall Street maintains a Strong Buy rating on the stock, with analysts projecting a 44% upside.
