Google has partnered with MediaTek to develop its next-generation AI chip, codenamed Triggerfish. This comes as a strategic shift in the company's custom silicon roadmap. With the new strategy, Google aims to strengthen its AI infrastructure while reducing dependence on a single chip design partner.
The partnership comes as technology companies continue to build specialized AI hardware to support increasingly complex generative AI models and enterprise workloads.
According to supply chain analyst Ming-Chi Kuo of TF International Securities, “ Triggerfish will feature a substantial memory upgrade, including SRAM capacity that is two to three times larger than Humufish and support for next-generation HBM4E memory, compared with HBM4 in the current design.”
Google has invested heavily in developing TPUs for its various applications and cloud platforms. Through its partnership with MediaTek, Google hopes to expand its semiconductor portfolio and accelerate chip development.
MediaTek is supposed to bring its chip design and system integration skills to Google. MediaTek also has strong relationships with TSMC, one of the companies that manufactures Google's custom AI processors.
This partnership can be seen as a big vote of confidence in MediaTek's expanding capabilities, not only in consumer electronics but also in other segments.
Google is developing its Triggerfish chip with MediaTek, packing up to three times the SRAM capacity of standard TPU v9 designs to handle the heavy processing demands of large-scale AI agents.
Custom silicon is now central to the AI arms race. Companies are building their own chips to cut dependence on third-party suppliers and lock in supply chains. For MediaTek, the partnership is a way into the AI infrastructure market.
Also read: Gemini Live Gets Context Upgrade as Google Deepens AI Personalization