Samsung Partners with Japanese AI Firm for 2nm Chip Solutions

Seoul, South Korea - Samsung Electronics (005930. KS) said it will provide 2-nanometer (nm) chip technology to Japanese artificial intelligence firm Preferred Networks, marking its first deal with a Japanese company.

In a press release, Samsung said it will supply turnkey solutions featuring its 2nm foundry process and 2.5D packaging technology, known as Interposer-Cube S (I-Cube S).

The collaboration aims to develop AI accelerators for generative AI applications requiring significant computing power.

Samsung's 2nm process uses a Gate-All-Around (GAA) transistor architecture, which the company claims improves performance and power efficiency. 

Taejoong Song, Samsung's head of Foundry Business Development, called the deal "pivotal" for validating the company's technology.

Samsung said the agreement represents a milestone for Japanese firms in heterogeneous integrated package technologies.

Preferred Networks, a major player in the global AI market, plans to use Samsung's technology to build energy-efficient, high-performance computing hardware.

The companies plan to showcase new AI chiplet solutions for data centers and generative AI computing markets in the future.

More Articles on “Samsung Foundry”

• Samsung's Q2 Profit Soars 1,452% on AI Chip Demand Surge
• Samsung Unveils Foundry Innovations and AI Strategy
• Samsung Reveals Automotive Strategy at SFF 2023
• Samsung's Game-Changing Automotive Tech Unveiled at IAA 
• Samsung Reveals Logic Semiconductor Advances