Trending Articles
Trending Topics
Trending
Nvidia is building out its range of AI tools targeting the autonomous vehicle sector.
The company announced Drive Alpamayo-R1 at this month’s NeurIPS artificial-intelligence conference, calling it “the world’s first open industry-scale reasoning vision language action model for mobility”.
According to Nvidia, the AI model will further the auto industry’s goal of creating fully autonomous vehicles, by enabling systems to think more like human drivers and use “common sense” when confronted with challenging driving situations.
Drive Alpamayo-R1 can break down elements of a driving scenario and then think through each element and possible route choices. It will use contextual data to weigh up each option up and pick the most appropriate one.
Drive Alpamayo-R1 and related data are now available under open-sources licences: Drive Alpamayo-R1 itself can be found on GitHub and Hugging Face, while its training data can be accessed via Nvidia Physical AI Open Datasets.
Drive Alpamayo-R1 is based on Nvidia’s reasoning vision language model Cosmos Reason. This is the foundation for a number of the company’s other autonomous-vehicle tools, including LidarGen, which can create lidar data for autonomous vehicle simulations, and Omniverse NuRec Fixer, which fixes artifacts in neurally generated data used for simulations.
According to research from analysts Precedence Research, the autonomous driving software market was worth $2 billion in 2025 and is forecast to reach over $7 billion by 2034.
It’s thought that the increase is being driven by a number of factors, including the increased use of AI in the sector – not only to help vehicles and their drivers identify and respond to hazards on the roads while driving, but also to help spot factors that cause damage to vehicles or increase their likelihood of breaking down.
Back in October, Nvidia announced a partnership with Uber to support the taxi company’s plans to create a network of 100,000 fully autonomous “robotaxis” and self-driving delivery vehicles. Uber will use Nvidia’s Drive software for the vehicles as well as its Drive AGX Hyperion 10 autonomous vehicle development platform.
Earlier this year, US-based electric vehicle maker Lucid announced that it aims to be the first car maker to sell level 4 cars to consumers (level 4 autonomous vehicles have the highest current level of driving autonomy). The
vehicles will make use of Nvidia’s multi-sensor suite architecture, as well as systems running on Nvidia’s Drive OS and Drive AGX Thor-powered hardware in an effort to achieve “eyes off, mind off” autonomous driving.
