IBM, Honda Developing AI Chips for Software-Defined Vehicles

IBM and Honda have inked a long-term research agreement to develop next-generation hardware technologies to power software-defined vehicles.

Honda and IBM said software-defined vehicles will “dramatically increase” the design complexity and power consumption of semiconductors.

The pair signed a Memorandum of Understanding and are set to jointly develop semiconductor solutions to increase processing performance while reducing power consumption when deployed in software-defined vehicles.

“Through this collaboration, the two companies would strive to realize software-defined vehicles that feature the world’s top-level computing and power-saving performance,” a statement read.

Together, IBM and Honda will work on semiconductors including hardware designed to emulate how the human brain processes information.

Brain-inspired AI, or neuromorphic computing, is a developing field in AI focused on building hardware that mimics the brain's ability to adapt and change through experience. Researchers believe this process enables computing systems to process information better.

IBM has already developed such semiconductors — its NorthPole hardware is designed to mimic how the human brain processes information on a single chip.

Honda and IBM will also work on chiplets, smaller and more modular hardware that can be deployed across different devices. A smaller piece of hardware deployed into the vehicle would not add unnecessary weight to the vehicle while still packing high-performance processing onto a smaller board.

Related:IBM Doubles Down on Enterprise AI With New Watsonx Assistants, Models

Additionally, the companies are researching related software solutions for software-defined vehicles.

“Hardware and software co-optimization is important to ensure high performance and fast time to market,” the companies said.

This article first appeared in AI Business' sister publication IoT World Today.