Samsung Electronics has announced a significant partnership with Preferred Networks, a leading Japanese AI company, to provide cutting-edge semiconductor solutions. The collaboration will utilize Samsung’s 2-nanometer (nm) foundry process and advanced 2.5D packaging technology, Interposer-Cube S (I-Cube S), to develop powerful AI accelerators.
This partnership marks a pivotal moment for Samsung, validating its 2nm Gate-All-Around (GAA) process technology and Advanced Package technology as ideal solutions for next-generation AI accelerators. Since initiating mass production of the industry’s first 3nm process node with GAA transistor architecture, Samsung has continued to strengthen its technological leadership in this field. The collaboration with Preferred Networks also represents a significant achievement in large-sized heterogeneous integrated package technologies for Japanese companies.
Preferred Networks, headquartered in Tokyo, is known for its vertical integration of the AI value chain, from chips to supercomputers and generative AI foundation models. The company has achieved notable success in the global AI market, securing the top position three times in the past five years on the Green500 list of supercomputers. By leveraging Samsung’s advanced foundry and packaging technologies, Preferred Networks aims to meet the increasing demand for computing power driven by generative AI applications.
Key Points:
- Samsung to provide 2nm foundry process and I-Cube S packaging technology to Preferred Networks
- Collaboration validates Samsung’s advanced technologies for next-generation AI accelerators
- Partnership marks first achievement for Japanese companies in large-sized heterogeneous integrated package technologies Preferred Networks aims to develop powerful AI accelerators to meet growing demand in generative AI
- Both companies plan to showcase groundbreaking AI chiplet solutions for data centers and generative AI computing in the future
