Japanese autonomous driving startup Tier IV has announced plans to develop an in‑house automotive AI chip, with its full design data to be released as open‑source – free for any automaker or tier‑1 supplier to use, modify, and fab. The goal: lower the barriers for carmakers venturing into self‑developed silicon and reduce dependence on overseas chip vendors.

The core of the project is a neural processing unit (NPU) purpose‑built for autonomous driving. Unlike general‑purpose processors, this NPU is optimized for AI workloads, delivering lower power per inference, and is designed to integrate with CPUs and other blocks on a single SoC. Tier IV will not manufacture the chips itself; instead, it will publish the complete design package in 2027, allowing others to customize and tape out via foundries – effectively skipping the multi‑year front‑end R&D phase and cutting project costs significantly.

Tier IV, founded in 2015 and backed by Toyota, Suzuki, and Isuzu, went public on the TSE Growth market in July 2026. Its core asset is the open‑source Autoware operating system, already widely used in autonomous driving R&D.

With Tesla, BYD, NIO, and NVIDIA all racing for dominance in automotive compute, Tier IV argues that proprietary chip development is prohibitively expensive and talent‑constrained. An open‑source silicon blueprint gives the entire industry a shared, customizable baseline – boosting supply chain flexibility and enabling more players to deploy sovereign vehicle‑grade compute.

ICgoodFind Takeaway:
Chip design costs keep many OEMs out of the game. Tier IV’s open‑source NPU could rewrite the rules – making custom automotive silicon accessible to all.