By ATN Editorial Team
Summary: Valeo has announced that it will develop and industrialise automotive-grade NVIDIA Hyperion compute platforms: high-performance compute platforms based on NVIDIA Hyperion, NVIDIA’s reference sensor and compute architecture for Level 4 autonomous vehicles and robotaxis, together with the NVIDIA Halos safety system. Valeo will apply its expertise in automotive integration, functional safety and industrialisation to make the platforms production-ready and tailored to automakers’ requirements.
Key engineering takeaway: The NVIDIA Hyperion compute platforms build on the dual NVIDIA DRIVE AGX Thor architecture within NVIDIA Hyperion, which Valeo says provides the advanced AI performance and redundancy required for Level 4 autonomous driving. The architecture is centralised, with a single platform intended to supply the compute for passenger vehicles, autonomous trucking and robotaxis.
Why it matters: Valeo frames the NVIDIA Hyperion compute platforms as a response to demand for centralised, high-performance compute in software-defined vehicles. By streamlining feature integration through a single-platform solution, Valeo says vehicle manufacturers can enhance operational efficiency, minimise system complexity and accelerate time-to-market.
Valeo, a global leader in mobility technology, today announced that it will develop and industrialize automotive-grade high-performance compute platforms based on the NVIDIA Hyperion reference architecture. This initiative combines Valeo’s automotive integration expertise with NVIDIA’s centralized compute architectures to accelerate the deployment of safe, scalable, and software-defined vehicle (SDV) systems.
As the automotive industry undergoes a fundamental shift toward software-defined mobility, the demand for centralized, high-performance compute power has become critical. Valeo is at the forefront of this transformation. Building on this experience, Valeo will develop high-performance compute platforms based on NVIDIA Hyperion, NVIDIA’s reference sensor and compute architecture for Level 4 autonomous vehicles and robotaxis along with NVIDIA Halos safety system. Valeo will apply its expertise in automotive integration, functional safety and industrialization to develop production-ready high-performance compute platforms based on this architecture and tailored to automakers’ requirements.
Dual NVIDIA DRIVE AGX Thor Architecture For Level 4
Building on the dual NVIDIA DRIVE AGX Thor architecture within NVIDIA Hyperion enables Valeo to provide the advanced AI performance and redundancy required for Level 4 autonomous driving. This unified architecture is engineered to support a wide spectrum of use cases, providing the centralized intelligence required for passenger vehicles, autonomous trucking, and robotaxis. By streamlining feature integration through this single-platform solution, vehicle manufacturers can enhance operational efficiency, minimize system complexity, and accelerate time-to-market for the next generation of intelligent transport.
“Our work with NVIDIA marks a significant milestone in our mission to make autonomous driving a safe, scalable reality,” said Marc Vrecko, CEO of Valeo Brain Division. “By seamlessly integrating NVIDIA’s advanced compute power into our next-generation platforms, we are delivering the flexible, centralized intelligence vehicle manufacturers require to redefine the driving experience and meet the evolving demands of tomorrow’s intelligent vehicles.”
“Valeo’s expertise in high-performance computing, vehicle integration and automotive-scale industrialization will help automakers adopt NVIDIA Hyperion and NVIDIA Halos at production scale,” said Rishi Dhall, Vice President of automotive at NVIDIA. “Together, NVIDIA and Valeo are providing the AI performance, functional safety and scalability needed to accelerate the development of Level 4 autonomous vehicles and robotaxis.”
NVIDIA Hyperion Compute Platforms: Frequently Asked Questions
What are Valeo’s NVIDIA Hyperion compute platforms?
They are automotive-grade high-performance compute platforms that Valeo will develop and industrialise based on the NVIDIA Hyperion reference architecture and the NVIDIA Halos safety system. Valeo will apply its expertise in automotive integration, functional safety and industrialisation to make them production-ready and tailored to automakers’ requirements.
What is NVIDIA Hyperion?
Valeo’s release describes NVIDIA Hyperion as NVIDIA’s reference sensor and compute architecture for Level 4 autonomous vehicles and robotaxis. In general terms, a reference architecture is a pre-defined combination of compute hardware, sensor set and supporting software that suppliers and vehicle manufacturers can build on, rather than designing every element of the system from scratch.
Why does Level 4 autonomous driving need redundant compute?
Valeo says the dual NVIDIA DRIVE AGX Thor architecture within NVIDIA Hyperion provides the advanced AI performance and redundancy required for Level 4 autonomous driving. In general, redundancy in automated driving compute means that a second processing path is available if one fails, so the system can keep operating or bring the vehicle to a safe state. This matters at Level 4, where the system rather than a human driver is expected to handle such situations within its operating domain.
What is Level 4 autonomous driving?
SAE Level 4 (high driving automation) describes vehicles that perform the entire driving task within a defined operational design domain, such as a mapped urban area or a highway route, without expecting a human driver to take over. Outside that domain the system does not operate.
Which vehicles are Valeo’s NVIDIA Hyperion compute platforms aimed at?
Valeo says the unified architecture is engineered to support a wide spectrum of use cases, providing the centralised intelligence required for passenger vehicles, autonomous trucking and robotaxis.
Further Reading
- Einride Adapts NVIDIA Hyperion Platform For Autonomous Trucking
- Valeo And KE:SAI Launch Open-Source Autonomous Driving AI
- Valeo Named System Tier-1 For Nissan Next-Generation ProPILOT
- What Is A Zonal Architecture? An Engineer’s Guide To Zone ECUs
For more autonomous vehicle news, click here.
