Samsung Electronics, a global leader in semiconductor technology, has announced a significant partnership with Preferred Networks, a prominent Japanese AI company. This collaboration will see Samsung providing turnkey semiconductor solutions utilizing its cutting-edge 2-nanometer (nm) foundry process and advanced 2.5D packaging technology, known as Interposer-Cube S (I-Cube S).
Preferred Networks aims to leverage Samsung’s state-of-the-art foundry and packaging technologies to develop high-performance AI accelerators. These accelerators are crucial in meeting the increasing demand for computing power driven by generative AI applications.
Samsung has been at the forefront of semiconductor innovation, notably being the first to mass-produce a 3nm process node with Gate-All-Around (GAA) transistor architecture. The company has further solidified its leadership by advancing to the 2nm process, promising enhanced performance and power efficiency.
This collaboration marks a milestone for Japanese companies in large-sized heterogeneous integrated package technologies. The 2.5D I-Cube S technology, essential for integrating multiple chips into a single package, boosts inter-connection speed and reduces package size, enhancing semiconductor performance.
GAONCHIPS, a specialized system semiconductor development company, designed the chip for this venture. This partnership not only underscores Samsung’s technological prowess but also aims to set new benchmarks in AI hardware development, particularly for generative AI and large language models.
Preferred Networks, headquartered in Tokyo, continues to lead in AI by vertically integrating AI value chains from chips to supercomputers. This partnership is poised to showcase groundbreaking AI chiplet solutions for future data centers and generative AI computing markets.
