Google is negotiating with Samsung Electronics to produce a component of its upcoming AI chip, Icefish, signaling a strategic move to expand manufacturing partners beyond TSMC and strengthen its AI hardware capabilities.
- Google’s Icefish AI chip aims for production by 2028
- Samsung could manufacture memory interface part with 2nm tech
- Move helps Google diversify manufacturing and Samsung grow contract chip business
What happened
Alphabet’s Google is reportedly in discussions with Samsung Electronics to manufacture a component of its next-generation AI processor known as Icefish. According to sources, Google intends TSMC to fabricate the processor’s main tensor processing unit, while Samsung would produce a chip section responsible for connecting the processor to memory using its advanced 2-nanometer fabrication process. The chip is currently under design development with potential mass production starting in 2028.
Google is collaborating with MediaTek on Icefish’s design, aiming to boost AI computing efficiency and power. Samsung declined to comment on the talks, and Google has yet to provide an official statement. This development comes amid Google’s efforts to expand its chip manufacturing partnerships beyond Taiwan Semiconductor Manufacturing Company (TSMC), the dominant foundry currently heavily engaged with AI demand.
Why it matters
This potential contract would represent a significant win for Samsung as it seeks to expand its footprint in the lucrative contract chip manufacturing market, particularly around highly advanced 2nm process technology. These smaller, more powerful chips offer improvements in speed, energy efficiency, and AI capabilities, positioning Samsung at the forefront of next-gen semiconductor manufacturing.
For Google, diversifying manufacturing sources reduces dependence on TSMC, which faces high demand and supply constraints from the booming AI hardware market. By involving Samsung, Google aims to secure a more resilient supply chain and accelerate production of its proprietary AI chips, which are key for competing against Nvidia and others in AI cloud services and hardware innovation.
What to watch next
Industry observers will be tracking whether Google finalizes manufacturing agreements with Samsung and how this affects production timelines for Icefish chips, expected to support advanced AI training and inference workloads. The collaboration’s progress may also impact Samsung’s expansion plans, including its prospects for additional semiconductor fabs such as a second Texas facility.
Concurrently, Google is reportedly in talks with Intel to produce millions of tensor processing units for 2028 as part of its broader strategy to diversify chip suppliers. Developments here could reshape competitive dynamics in AI chip production and influence supplier partnerships in the semiconductor industry over the coming years.