Samsung updated its semiconductor process roadmap at its foundry forum in San Jose, targeting mass-production of a fourth generation 2nm chip in 2027 and an upgraded 4nm silicon in 2025. 

At the Samsung Foundry Forum 2024, the company stated its latest 2nm process uses advanced backside power delivery network technology to improve power efficiency and lower temperatures. The SF2Z process node will be ready for mass production in 2027.

The new SF4U process, referred to as a high value 4nm variant, offers power, performance and area improvements using “optical shrink”, with mass production scheduled for 2025, the company said.

Choi Si-young, president of the company’s foundry business (pictured), stated high-performance, low-power semiconductors are key at a time when numerous technologies are evolving around AI.

He noted Samsung plans to introduce integrated, co-packaged optics technology for high-speed, low-power data processing.

The company also claimed its gate-all-around (GAA) transistor architecture is maturing in performance and yield.

Samsung noted preparations for the 1.4nm process (SF1.4) are “progressing smoothly, with performance and yield targets on track”, adding it is “actively shaping” future process technologies below 1.4nm through material and structural innovations.

Its first-generation 2nm process, SF2, will begin in 2025, with an upgraded SF2P version to be ready in 2026.

Data from TrendForce showed Samsung’s share of the foundry market in Q1 was flat year-on-year at 11 per cent, while rival Taiwan Semiconductor Manufacturing Company held 61.7 per cent.

Samsung was awarded a 2nm AI chip production contract from Japanese start-up Preferred Networks in February.

Last month, the chipmaker replaced the head of its semiconductor business as it pushes to gain on rival SK Hynix in the high bandwidth memory chip segment.