Bos Semiconductor recently announced on the 11th that it has signed a development contract for chiplet semiconductors for next-generation autonomous driving systems with a major automotive company in Europe, specifically for an advanced driver-assistance system (ADAS).
Through this collaboration, joint development will be conducted throughout the entire process, from architectural design of the semiconductor to vehicle system verification and ADAS AI software optimization. This contract aims to configure Bos Semiconductor's Eagle-N (automotive chiplet AI accelerator) and Eagle-A (standalone ADAS semiconductor) as a chiplet system for use in customers' vehicle next-generation systems.
Park Jae-hong, CEO of Bos Semiconductor, noted that "this ADAS chiplet semiconductor development contract with a major leading customer is a significant achievement for Bos Semiconductor's product mass production." He emphasized that the Eagle chiplet semiconductor product line from Bos Semiconductor is competitive in performance, expense efficiency, power efficiency, AI software optimization, and system software flexibility to be used in all ranges of vehicle ADAS systems from low-end to mid-range and high-end.
In the automotive industry, ADAS solutions have become a key element in enhancing vehicle safety and supporting autonomous driving technology. In particular, as autonomous driving technology advances to level 2 and beyond, the demand for high-performance, energy-efficient semiconductors for real-time data processing is skyrocketing. However, the existing monolithic package method has limitations in scalability, flexibility, cost competition, and performance optimization.
In contrast, chiplet-based architecture, with its modular design, allows for the use of multiple small, low-cost chips tailored to specific ADAS functions, emerging as a solution that overcomes the problems of single semiconductor packages. The Eagle-N is Bos Semiconductor's first product, the industry's first automotive chiplet AI accelerator semiconductor, boasting up to 250 TOPS of neural processing unit (NPU) performance. Discussions for sample verification are currently underway with several global clients, targeting mass production in 2026.
The next product, Eagle-A, is a standalone ADAS semiconductor capable of integrating sensing from camera/LiDAR/Radar, equipped with a powerful AI NPU processor and high-performance CPU/GPU. Sample release is aimed for the first quarter of 2027, with mass production targeted for 2028. The Eagle ADAS chiplet semiconductor product line can be mounted in various vehicles as a flexible and scalable system by combining one Eagle-A with multiple Eagle-Ns to implement autonomous driving levels 2 to 4.