loading
hello dummmy text
Subaru AI system development with HPE servers

Subaru is stepping up its AI development game by selecting Hewlett Packard Enterprise’s (HPE) Cray XD670 servers to power the next evolution of its EyeSight driver assistance technology. This move represents a major investment in accelerating the design and refinement of AI models that underpin Subaru’s safety systems.

The EyeSight system, already known for its proven effectiveness in reducing collisions, is undergoing a significant upgrade. Subaru’s upcoming version aims to combine stereo camera vision with advanced artificial intelligence to accurately measure distances and classify objects in real-time — a crucial step toward making driving safer under a wider range of real-world conditions.

To bring this ambitious project to life, Subaru has chosen to deploy the HPE Cray XD670 servers within its own facilities. This on-site setup gives engineers full control over a high-performance computing environment optimized for AI workloads. The new system is built with eight Nvidia H200 GPUs and utilizes direct liquid cooling to deliver high efficiency and top-tier performance while managing energy consumption.

Takashi Kanai, Deputy Chief of Subaru Lab and General Manager of the company’s ADAS development division, emphasized the impact of this upgrade. “By integrating HPE Cray XD670 servers into our infrastructure, we’re doubling our computational performance compared to our previous system. This enables faster training and fine-tuning of AI models, helping us speed up the entire development lifecycle.”

Subaru engineers are collecting extensive road video data by testing vehicles in various global environments. These videos are used to develop and test machine learning models that can predict and respond to complex driving scenarios. With the new computing infrastructure, this data can now be processed and analyzed much faster, significantly improving development timelines.

The decision to go with HPE was driven not just by performance but also by reliability and scalability. The Cray XD670 platform is known for supporting demanding applications such as computer vision and deep learning — both essential components of Subaru’s new AI-based safety systems.

HPE Japan’s Senior Vice President and Managing Director, Hirokazu Mochizuki, expressed pride in the collaboration. “Subaru’s commitment to innovation and driver safety aligns perfectly with our vision for high-performance computing. We’re excited to support their mission to eliminate fatal road accidents through smarter technology.”

EyeSight plays a key role in Subaru’s broader “All-Around Safety” approach, which focuses on proactive and preventive technologies in addition to traditional safety systems. This includes areas such as driver monitoring, collision avoidance, and connected services.

By combining deep AI expertise with cutting-edge computing hardware, Subaru is reinforcing its long-term vision: creating vehicles that deliver safety, peace of mind, and advanced technological performance — all while accelerating toward a future with fewer accidents on the road.

Write a Reply or Comment

Your email address will not be published. Required fields are marked *

Recent Posts