Nvidia and Uber Join Forces to Accelerate Autonomous Driving
Chipmaker Nvidia and ride-hailing giant Uber are teaming up to advance autonomous driving technology, a move that sent Uber’s stock up 3.5% on Thursday afternoon.
According to a post on X, the partnership centers on Nvidia’s Cosmos World foundational artificial intelligence (AI) model, which will now be trained using Uber’s vast repository of real-world driving data including airport pickups, complex intersections, and variable weather conditions. By feeding that data into Cosmos, Nvidia aims to enhance the model’s ability to simulate and reason through unpredictable situations, shortening the testing cycle and improving performance in rare or extreme driving scenarios. The collaboration relies on Nvidia’s DGX Cloud infrastructure and targets three key goals: achieving higher precision in simulation, speeding up post-training iterations, and ensuring more reliable model behavior in challenging conditions.
The effort builds on Nvidia’s broader roadmap for AI-enabled driving, described in an Oct. 20 company blog post, “How AI Is Unlocking Level 4 Autonomous Driving.” The post outlines how a series of AI advances are converging to make high automation commercially viable. Foundation models, the company explains, can draw on internet-scale knowledge rather than limited driving datasets, allowing vehicles to generalize from vast training data. End-to-end architectures, meanwhile, allow a single network to process sensor inputs directly into driving decisions without losing context, improving performance and reducing engineering complexity.
Simulation is another pillar of Nvidia’s approach. Technologies such as its Cosmos Predict and Transfer systems can generate new weather, lighting, and traffic conditions on demand, enabling autonomous vehicles to “practice” millions of edge cases virtually before encountering them on real roads. These capabilities are supported by Nvidia’s DRIVE and DGX compute platforms, which train, test, and deploy AI driving models from the cloud to the car.
“With foundation models, a vehicle encountering a mattress in the road or a ball rolling into the street can now reason its way through scenarios it has never seen before, drawing on information learned from vast training datasets,” Nvidia wrote in the post.
The collaboration extends a growing convergence between the AI and mobility sectors. PYMNTS has previously reported on Nvidia’s push to embed AI infrastructure into automotive systems and on Uber’s data-driven evolution of its logistics and mobility platforms, a combination that could accelerate the path toward scalable Level 4 autonomy.
The post Nvidia and Uber Join Forces to Accelerate Autonomous Driving appeared first on PYMNTS.com.