Google is in talks with Samsung to produce a component for its tenth-generation Tensor Processing Unit, codenamed Icefish. Samsung is reportedly set to use its 2nm technology for a memory input-output unit. TSMC will remain responsible for the computing engine.
This is reported by The Information, citing sources familiar with the matter. This potential move is no coincidence. TSMC, which Google has historically relied on for its chip production, is facing capacity shortages due to high demand for AI chips, particularly from Nvidia. According to recent analyses, TSMC produces more than 90 percent of the world’s advanced AI chips. This is putting pressure on available production capacity.
In the current design, TSMC remains responsible for the most demanding part of the chip: the computing engine, manufactured using 1.4nm technology. Samsung would produce the component that connects the chip to the memory.
For the Icefish project, Google is also collaborating with the Taiwanese chip designer MediaTek.
Samsung’s 2nm technology on the rise
Samsung is expanding its 2nm production capacity, but its share of the advanced AI chip market remains limited. TSMC completely dominates this segment. The deal with Google would represent a significant step toward Samsung’s ambitions to attract more major AI chip customers for its foundry division.
It is not known exactly when a potential final agreement will be reached. Mass production of Icefish is currently scheduled for 2028.