TIER IV and TUM Bring NVIDIA Hyperion to Autoware Projects in Germany
TIER IV and the Technical University of Munich (TUM) are expanding their collaboration to bring NVIDIA Hyperion and the NVIDIA Halos safety system into Autoware-based autonomous vehicle projects across Germany. The initiative was announced at NVIDIA GTC Berlin and targets Level 4 autonomous driving.
- TIER IV and TUM announced an expanded collaboration at NVIDIA GTC Berlin.
- NVIDIA Hyperion, including dual-DRIVE AGX Thor configurations, will be integrated into Autoware-based projects across Germany.
- The work builds on the recently established TUM Autoware Research Center by TIER IV.
- Focus areas are autonomous public transportation and robotaxi applications.
- No timeline, locations or performance figures were disclosed.
What the partners announced
TIER IV, which describes itself as the force behind Autoware, the world's leading open-source software for autonomous driving, and TUM said they will build on the recently established TUM Autoware Research Center by TIER IV. The new initiative will integrate NVIDIA Hyperion, including high-performance dual-DRIVE AGX Thor configurations, into Autoware-based autonomous vehicle projects in Germany.
According to the announcement, NVIDIA Hyperion is a reference sensor and compute platform for Level 4 autonomous vehicles and robotaxis. It combines NVIDIA DRIVE AGX Thor accelerated computing with the NVIDIA Halos safety system, with the aim of accelerating safe, production-ready autonomy.
Goal: from research to real-world deployment
The partners say that bringing together NVIDIA's automotive-grade compute and safety technologies, Autoware's open-source software and TUM's automotive research expertise will give researchers and developers an environment for building, testing and validating next-generation autonomous driving systems.
The work will focus on translating advanced research into safe, scalable real-world deployments. The release names autonomous public transportation and robotaxi applications as examples.
Prof. Markus Lienkamp of TUM said the partnership allows the team to "rigorously test autonomous vehicle systems on German roads," with public transit solutions built on "transparent, verifiably safe, and collaborative technology rather than restrictive proprietary systems." Prof. Johannes Betz said that years of joint work on Autoware between TUM and TIER IV, combined with platforms like NVIDIA Hyperion, lets the team push real-time vehicle control and software architecture "to a whole new level."
The open-source argument
Both the university and TIER IV stress openness. Shinpei Kato, founder and CEO of TIER IV, said that scaling safe autonomous driving globally requires "moving beyond closed, black box systems toward open, interoperable architectures." He called TUM an exceptional research partner and said NVIDIA Hyperion, together with NVIDIA's compute leadership and its commitment to open-source software ecosystems, provides the foundation needed to accelerate scalable, publicly trusted autonomous transit in Germany and beyond.
Rishi Dhall, vice president of automotive at NVIDIA, said Hyperion provides the reference sensor and compute architecture needed to develop Level 4 autonomous vehicles and robotaxis at scale. He added that the three parties are combining DRIVE AGX Thor and Halos with an open-source autonomous driving platform.
What it means
For researchers and developers working with Autoware in Germany, the initiative promises access to a reference hardware and safety stack alongside the open-source software they already use, with testing on German roads. The announcement does not give a timeline, vehicle counts, pilot locations or performance figures, so the scale and pace of deployments remain to be seen.
