After causing a veritable maelstrom in the memory sphere by hogging the global DRAM capacity for AI-focused high bandwidth memory (HBM), to the detriment of all other DRAM uses, the ongoing AI mania is now on the cusp of creating similar dynamics in the mobile chips sphere, judging from MediaTek's decision to deprioritize its mobile SoC division by diverting resources towards AI ASICs, which might reduce the competitiveness of MediaTek's Dimensity chips further down the line.
MediaTek has redirected internal resources away from its mobile chips division and towards blue ocean markets such as AI ASICs and automotive silicon
According to Taiwan's CTEE, MediaTek has moved some of its resources, including personnel, from its mobile chips division to blue ocean products such as AI ASICs and automotive silicon. For the benefit of those who might not be aware, a blue ocean market is one where opportunities abound, and competition is either limited or nonexistent.
We noted recently that MediaTek had played an important role in the development of Google's TPU v7 Ironwood ASIC by designing its input/output (I/O) modules that facilitate communication between the processor and peripherals. This constituted a departure from Google's typical strategy in recent years, where it designed the entirety of next-gen TPU in a close collaboration with Broadcom.
MediaTek intends to deepen its ASIC collaboration with Google as the next-gen TPU is all set to enter the volume production stage in Q3 2026, with Google intending to manufacture 5 million units of its bespoke ASIC in 2027 and 7 million units in 2028.
To prepare for this onslaught of volume, both Broadcom and MediaTek have reportedly increased their wafer starts. Of course, the overall complexity of Google's TPU has also increased now that it leverages TSMC's 3nm process. As such, MediaTek felt compelled to divert resources away from its mobile chips division to create a dedicated team for its ASIC-related ambitions.
MediaTek is banking on its proprietary SerDes (serializer/deserializer) technology - one that converts parallel data into a high-speed serial stream for efficient transmission, only to convert it back to parallel at the receiver - to gain an edge in the ASIC sphere. This tech allows for speedy and efficient communication between the processor and the memory, which significantly enhances the potency of a given AI ASIC.
MediaTek's current-gen 112Gb/s SerDes DSP uses a PAM-4 receiver architecture and manages to achieve over 52dB of loss compensation capability at a 4nm process while maintaining low signal attenuation and high anti-interference characteristics, which are critical incentives for data centers and advanced packaging architectures. The chip designer is also working on its next-gen 224Gb/s SerDes DSP.
MediaTek expects to earn $1 billion from AI ASICs in 2026, scaling to "several billion dollars" in 2027. In addition to Google's TPU, the chip designer is also trying to rope in Meta for an expansive collaboration on its bespoke ASIC. According to industry insiders, MediaTek now sees AI as a structural transformation of its growth engine, one that entails an apparent deprioritization of its mobile chips division.
Of course, MediaTek's Dimensity chips remain eminently competitive for now. As we noted recently, both Qualcomm and MediaTek have switched to TSMC's N2P process for their upcoming flagship chips, enabling higher clock speeds with less of an efficiency-related drag. However, now that MediaTek is actively deprioritizing its mobile chips division, we are unsure how long the Dimensity chips will remain competitive, especially as these chips were never the apex predator in the mobile SoC arena to begin with, where Apple's A-series chips and Qualcomm's Snapdragon ones have reigned supreme for a long time.
About the author: Writing is my one incontrovertible passion. Over the past six years, he has authored over 2,200 distinct articles on financial and tech-related topics, spanning nearly 1 million words. And he has been a member of Wcctech mobile team since 2025. As an alumnus of the University of Toronto, Rotman Commerce Program, I bring nuance, in-depth knowledge, and a unique perspective to every topic that I cover. When I'm not writing, I'm traveling the world, exploring hidden confectionaries and restaurants as an aspiring food connoisseur.
Follow Wccftech on Google to get more of our news coverage in your feeds.
