Mercedes-Benz and Nvidia announced an extension of their technology partnership on Oct. 16, aiming to embed the chipmaker’s compute and sensor solutions into the automaker’s upcoming models that run the MB.OS operating system.
Mercedes-Benz to use Nvidia’s Hyperion platform
Future vehicles from the German brand will incorporate the firm’s Hyperion in-vehicle compute platform. The system relies on dual Nvidia Drive AGX Thor high-performance processors, which enable parallel processing of sensor data and support higher-resolution perception. A standardized sensor suite works with those processors to deliver real-time analysis for advanced driver assistance and autonomous functions.
Hyperion also supports over-the-air updates, allowing software improvements without dealer visits. Those updates can add new features, patch security gaps, and refine algorithms, thereby enhancing safety and convenience throughout the vehicle’s life cycle.
Scaling Compute for AI Model Development and Validation
Advanced autonomous driving systems now need massive compute capacity. The requirement covers training AI models, simulating driving behavior, validating performance, and handling real-time processing inside the vehicle. Such processing power is essential for delivering reliable Level 4 performance.
Hyperion enables continuous, large-scale data collection from operational cars, feeding ongoing AI model training to refine autonomous capabilities. By pairing Hyperion with Nvidia’s Alpamayo stack, Mercedes-Benz gains full-stack reasoning models and software that speed autonomous vehicle and robotaxi development. Alpamayo also provides closed-loop testing of ADAS and other autonomous functions within a safe simulated environment. This approach allows iterative improvement without requiring physical test drives for every software update.
Software stack and safety systems
Alongside the hardware, the automaker will adopt the chipmaker’s open-source Alpamayo full-stack software platform. Alpamayo provides AI reasoning capabilities that aim to make complex driving decisions similar to those of a human driver, and it supports sensor-fusion techniques that improve object detection accuracy. The stack targets SAE Level 4 automation, where the vehicle can handle most driving situations without driver intervention.
The company will also integrate the firm’s Halos safety system, which provides system vehicle software safeguards and redundancy of critical vehicle sensors. Halos offers validation tools throughout the vehicle development cycle, helping engineers verify that safety requirements are consistently met.
“Nvidia is our most important strategic partner in automated driving,” said Jörg Burzer, Mercedes-Benz Group Chief Technology Officer. “The expansion of our partnership marks a significant milestone and reflects the mutual trust we have built over the years.”
Ali Kani, vice president and general manager of the chipmaker’s automotive team, added, “Accelerated computing is giving vehicles the intelligence to understand the world and reason through complex driving situations.”
The chipmaker’s role as a hardware supplier has expanded beyond the brand. General Motors, Hyundai Motor Group, and Volvo Cars also use its processors and AI-driven software in their vehicle programs, illustrating a growing ecosystem of partners.
In March 2025, General Motors announced a partnership with the firm to equip its future Cadillac Escalade IQ electric SUV with a “hands-off” driving capability built on the Drive AGX platform.
Earlier this year, Hyundai Motor Group disclosed an expanded agreement with the chipmaker to speed the rollout of Level 2 driver-assist features and lay groundwork for Level 4 autonomous technology.
The company also employs the firm’s engineering simulation tools within its Mercedes-Benz Cars Operations 360 data platform. The digital-twin capability helps the automaker optimize assembly processes, reduce material waste, and improve quality across its global production network.