Samsung foundry eyes key role in Google's next-gen TPU production

Samsung Electronics may be tapped to produce a key component of Google's next-generation AI chip, the tensor processing unit (TPU), according to industry observers. The Taylor fab in Austin, Texas — where Samsung is ramping up new investment — is seen as the likely production base.

Google has been pursuing a strategy to diversify its TPU production as a bottleneck has emerged at TSMC, where orders from major technology companies have piled up. Reports of Intel Foundry winning orders have fueled growing interest in whether Samsung Electronics' foundry division could also expand its footprint with Google. According to The Information, a US technology publication, Google is in discussions to hand production of a key component of its next-generation TPU to Samsung Electronics' foundry division.

Google is developing its 10th-generation TPU under the code name Icefish. The chip is being designed in collaboration with Taiwanese semiconductor design firm MediaTek, with mass production expected to begin as early as 2028. The TPU is an AI chip of Google's own making and has been central to the success of its Gemini AI platform.

Google plans to split production of the 10th-generation TPU's components between Samsung Electronics' foundry and TSMC. The main processor handling computation would use TSMC's 1.4-nanometer process, while the memory input/output die — which connects the processor to HBM — would be produced using Samsung's 2nm process.

Google is said to have placed high value on Samsung Electronics' deep understanding of memory, given its status as an integrated device manufacturer (IDM). Google is believed to have factored in Samsung's grasp of the characteristics and specifications of memory chips, including HBM. Samsung is the only company capable of offering a one-stop turnkey solution spanning memory, foundry, system LSI and advanced packaging. Samsung also supplies the HBM used in Google's TPUs and is locked in a fierce competition with SK Hynix for the position of primary supplier. Some in the industry project that Samsung's HBM supply volume to Google will surpass SK Hynix's starting this year.

Reports have also emerged that Google plans to award Intel Foundry production of more than 3 million TPUs for use from 2028 onward, after evaluating Intel's packaging technology — a volume that would account for roughly half of total production over that period.

TSMC has long held a monopoly on TPU production, but demand for AI chips from Google, Nvidia, Amazon, Microsoft and other major technology companies has surged, creating a production bottleneck. Factories running at full capacity are still unable to keep pace with demand.

As Google turns its attention to Intel, Samsung Electronics' foundry division is also being mentioned as a potential beneficiary.

Samsung plans to meet demand from US technology companies by scaling up its Taylor fab in Austin, Texas. The memory I/O die for Google's 10th-generation TPU is expected to be produced using a 2nm process, and Samsung intends to center its Taylor fab strategy around that node.