Alphabet’s Google is in discussions to collaborate with Taiwan’s MediaTek for the upcoming generation of its AI chips, known as Tensor Processing Units, with production set for next year, according to a report from the Information that cites sources familiar with the matter.
Despite this new partnership, Google has maintained its relationship with Broadcom, the chip maker with which it has exclusively collaborated on AI chips for several years, as noted by an employee from the San Jose-based company.
Requests for comments from Google, MediaTek, and Broadcom were not met with immediate responses from Reuters.
In a similar vein to Nvidia, Google also engineers its own AI server chips, employing them for internal research and development while also offering them to cloud clients.
This strategy provides Google with a competitive advantage in the AI sector, diminishing its dependence on Nvidia, even amid rising demand for Nvidia chips from competitors such as Microsoft-backed OpenAI and Meta Platforms.
Last year, Google introduced its sixth-generation TPU, aiming to present both itself and its Cloud customers with an alternative to Nvidia’s highly coveted processors.
The decision to partner with MediaTek was influenced by the Taiwanese company’s strong ties with TSMC, coupled with its lower cost per chip compared to Broadcom, the report indicated.
According to the research firm Omdia, Google invested between $6 billion (approximately Rs. 51,952 crore) and $9 billion (around Rs. 77,929 crore) in TPUs last year, which aligns with Broadcom’s predicted revenue from AI semiconductors.
© Thomson Reuters 2025
(This story has not been edited by NDTV staff and is auto-generated from a syndicated feed.)
