#Japan #Tokyo #NVIDIA #TIER IV #Autoware

Overview

Tier IV, headquartered in Shinagawa, Tokyo, is on a mission to democratize autonomous driving. The company is intensifying its collaboration with NVIDIA Corporation to accelerate the development of Level 4 autonomous driving technology. With a vision for a safer and more scalable future, Tier IV is focused on integrating NVIDIA's reasoning-based self-driving AI models alongside their open-source software platform, Autoware.

Collaboration with NVIDIA

Under the guidance of CEO Shohei Kato, Tier IV is leveraging NVIDIA’s VLA (Vision-Language-Action) model and other tools, such as NVIDIA Alpamayo, to enhance their autonomous driving capabilities. This collaboration aims to enrich Autoware with advanced reasoning layers, allowing for more nuanced decision-making in complex scenarios.

Integration of NVIDIA Alpamayo

The integration of the NVIDIA Alpamayo model marks a significant leap for Tier IV's development efforts. This model, featuring 100 billion parameters, facilitates enhanced interpretation of intricate traffic situations through its reasoning capabilities. By employing a cascade of thought processes, the software can analyze and respond to unpredictable real-world scenarios, thus augmenting the transparency and accuracy of its decision-making processes. The company aims to incorporate the latest models unveiled at the NVIDIA GTC 2026 event into their platform, reinforcing their commitment to safe, commercial scalability for autonomous driving.

Enhancing Co-MLOps Platform with NVIDIA Cosmos

In an effort to speed up the development of autonomous driving AI, Tier IV plans to roll out their Co-MLOps platform in 2024, designed to facilitate extensive data sharing across its global partnerships. Utilizing NVIDIA Cosmos, the platform will tackle challenging edge cases that traditional learning methods struggle to manage. Key features of the Co-MLOps platform include:

• - Cosmos-Predict: This function enables the generation of edge cases through multimodal instructions, producing accurate synthetic data for rare real-world situations.
• - Cosmos-Transfer: Utilizing self-labeling foundations, this feature can adapt label images to various environmental conditions, enhancing data diversity through transformations like adapting to heavy rain, snow, or nighttime.
• - Cosmos-Reason: This function allows for rapid searching, verification, and summarization of essential information derived from vast driving datasets using a visual language model that captures the essence of the physical world.

Comments from Leadership

Shohei Kato, CEO of Tier IV, emphasized the need for a data-centric AI model that can address unpredictable real-world events. By adopting NVIDIA Alpamayo and integrating NVIDIA Cosmos into the Co-MLOps platform, they’re contributing to the Autoware community's vision of achieving safe and scalable autonomous driving technologies.

Marco Pavone, Research Director for Autonomous Vehicles at NVIDIA, noted that physical AI represents a core technology redefining the landscape of AI. He underscored how Tier IV is maximizing their capabilities with NVIDIA Alpamayo and Cosmos, setting advanced precedents within their ecosystem for creating safe and transparent autonomous systems.

Participation in NVIDIA GTC 2026

Tier IV's involvement in the NVIDIA GTC 2026 event in San Jose, California focuses on showcasing their Co-MLOps platform and the application of NVIDIA Cosmos within their autonomous driving data set infrastructure. The session aims to illustrate how these advancements contribute to the future of autonomous vehicle technology.

About Tier IV

Founded in December 2015, Tier IV is a deep tech company leading the development of the open-source autonomous driving software, Autoware. They strive to expand the possibilities of autonomous driving while collaborating with global partners to create a safer and sustainable society. Their offerings include development services for autonomous driving platforms, web services, system development kits, and educational initiatives in autonomous technologies.

For more information, visit Tier IV.