October 20, 2026 -
BOS Semiconductors, a fabless semiconductor company specializing in Physical AI and automotive semiconductors, has joined forces with LG Electronics, HANA Micron, and other partners to advance the domestic development and commercialization of chiplet-based automotive semiconductors.
On September 30, BOS Semiconductors signed a Memorandum of Understanding (MOU) for cooperation on the domestic development of chiplet-based automotive semiconductors with LG Electronics, HANA Micron, and four organizations specializing in automotive and industrial technologies. The agreement was signed during the "Future Mobility Ecosystem Cooperation" event hosted by Korea's Ministry of Trade, Industry and Resources (MOTIR).
The agreement is part of the M.AX Alliance (Manufacturing.AX Alliance) initiative led by MOTIR. It aims to bring together the capabilities of Korean companies and institutions to establish a collaborative ecosystem covering the entire value chain of next-generation automotive semiconductors, from development and integration to packaging, testing, validation, and commercialization.
Under the agreement, BOS Semiconductors will develop and supply Neural Processing Unit (NPU) chiplets designed to support autonomous driving and in-vehicle AI computing. The company will also contribute to chiplet interoperability validation, optimization, and commercialization. LG Electronics will provide a base chiplet responsible for core automotive system functions, while HANA Micron will leverage its semiconductor packaging and testing expertise to develop multi-chiplet packaging solutions and provide performance validation and technical support. Together, the partners aim to establish a domestic automotive chiplet value chain encompassing AI semiconductor design, chiplet integration, advanced packaging, testing, validation, and real-vehicle testing.
Chiplet technology represents an alternative to integrating all functions into a single large semiconductor chip. Instead, individual functions such as AI processing, CPUs, and I/O are designed as separate chiplets and connected through high-speed interconnects to operate as a unified system. This modular approach enables flexible combinations of chiplets based on specific performance and functional requirements, while reducing development time and costs through the reuse of existing chiplets. In the automotive industry, where computing requirements vary by vehicle segment and the level of autonomous driving and AI functionality, chiplet architectures offer the ability to scale performance on a common platform, improving development efficiency and cost competitiveness.
BOS Semiconductors has long recognized chiplet technology as a key architecture for next-generation automotive AI semiconductors and has continued to develop its capabilities in this area. Building on its integrated AI SoC, Eagle-A, and high-performance AI accelerator, Eagle-N, the company is developing a chiplet-based automotive computing platform that optimizes system processing and AI computing functions separately while enabling flexible scaling of AI computing resources according to product requirements. Rather than relying on a single fixed SoC architecture to address diverse vehicles and applications, BOS Semiconductors aims to deliver a modular and scalable architecture that can be configured and expanded to meet specific performance and functional needs.
BOS Semiconductors is also actively participating in the ecosystem surrounding Universal Chiplet Interconnect Express (UCIe), a global standard for chiplet interconnects. UCIe enables high-speed communication between different chiplets, allowing them to operate as an integrated system and serving as a key technology for achieving chiplet interoperability.
In addition to its participation in the UCIe Consortium, BOS Semiconductors is engaged in the Open Chiplet Atlas (OCA) ecosystem, contributing to technology development focused on chiplet interoperability. More recently, leveraging its expertise in automotive chiplet technologies, the company was selected as a technology partner for MIRISE Technologies' automotive chiplet interconnect research and prototyping project.
"As automotive systems evolve, architectures that combine semiconductors optimized for specific functions and allow AI computing performance to scale flexibly will become increasingly important, alongside traditional approaches that integrate all functions into a single large SoC," said Jason Chae, Executive Vice President and Head of Strategic Marketing at BOS Semiconductors. "BOS Semiconductors has continued to expand its capabilities not only in NPU chiplets, but also in UCIe-based interconnect and system integration technologies essential for implementing automotive chiplet solutions."
"Through this collaboration, we aim to connect our chiplet technologies with the broader automotive semiconductor ecosystem in Korea and globally, while expanding partnerships that can lead to future product development and commercialization," he added.