Samsung Foundry unveiled its updated semiconductor chip process roadmap today. During the ongoing SFF 2024 expo in the US, the South Korean semiconductor chip giant unveiled its process roadmap for 2nm and 3nm chips and plans for specialized nodes for AI and automotive chips.
Samsung Foundry plans to make 2nm chips in 2025 and 1.4nm chips in 2027
Today, Samsung held the Samsung Foundry Forum 2024 annual expo at its Device Solutions America headquarters in San Jose, California. During the event, it unveiled its updated process technology roadmap, which includes two new cutting-edge nodes: SF2Z and SF4U.
The company revealed that it will be ready to start mass-producing 2nm chips for mobile devices in 2025. Its first-generation 2nm process, called SF2, will be ready next year, and the improved version of its 2nm process, called SF2P, will be ready in 2026.
A specialized version of its 2nm node, SF2X, is designed for Artificial Intelligence (AI) and High-Performance Computing (HPC and servers) chips and will be ready in 2026.
The company's fourth-generation 2nm process node, SF2Z, uses the advanced backside power delivery network (BSPDN) technology for even better power efficiency and temperatures. It uses power rails on the backside of the wafer to remove bottlenecks between the power and signal lines. It will be ready for mass production in 2027.
A variant of its 2nm process for automobile chips, SF2A, will also be ready for mass production in 2027. The company claimed its Gate All Around (GAA) transistor architecture is maturing continuously in both performance and yield. GAA, which was introduced with its 3nm process, will also be applied to its 2nm processes.
Samsung started mass-producing 3nm (SF3E) chips in the second half of 2022, but it could only get contracts for relatively simpler cryptocurrency mining chips. Its second-generation 3nm process is called SF3 and is ready for mass production. While the company didn't reveal information about possible chips using this technology, some reports say Samsung's own Exynos W1000 will be the first chip to use this technology.
Later this year, the company could also start mass production of its next-generation Exynos smartphone chipset for the Galaxy S25 series. It could also use the SF3 process and could be unveiled early next year.
SF4X, the company's fourth-generation 4nm process node, is ready for AI and HPC use. Next year, the company will be ready with SF4A, a variant of its 4nm process node optimized for automotive chips. SF4U, its brand-new 4nm process node, offers performance, power, and area (PPA) improvements using optical shrink. It will be ready for adoption in 2025.
Dr. Siyoung Choi, President and Head of Samsung Foundry, said, “At a time when numerous technologies are evolving around AI, the key to its implementation lies in high-performance, low-power semiconductors. Alongside our proven GAA process optimized for AI chips, we plan to introduce integrated, co-packaged optics (CPO) technology for high-speed, low-power data processing, providing our customers with the one-stop AI solutions they need to thrive in this transformative era.”
The story continues after the video.
Samsung also unveiled its integrated Samsung AI Solutions platform, which helps clients by offering Samsung's unique strengths, which are advanced packaging, foundry, and memory. Its rival, TSMC, lacks expertise with semiconductor memory chips.
Harnessing its unique strengths, the company plans to offer high-bandwidth, high-performance, and low-power solutions that can be customized to customers' AI requirements. With its turnkey AI solution, the company plans to attract clients like AMD and Nvidia.
Samsung Foundry is also trying to diversify its client portfolio by catering to semiconductor chip needs in the Internet of Things (IoT), medical, and wearable segments. It is partnering with several chip design and testing partners, including ARM and Groq, to create a semiconductor ecosystem that progresses faster.
