Google teams up with Mediatek to develop next-gen AI chips

Nvidia has long had a chokehold on the AI chip market with most companies, including Meta, Google, and OpenAI, heavily reliant on its processors. However, that could soon change as Google partners with MediaTek to develop its own AI chips.

Reports suggest Google is working with Taiwan’s MediaTek to produce the next generation of Tensor Processing Units (TPUs)—its custom AI chips used for cloud services and AI applications. These chips are expected to enter production next year, potentially giving Google an edge in AI-powered devices and services.

While Google has traditionally partnered with Broadcom for chip development, this new collaboration with MediaTek signals its intent to diversify suppliers. The move is also driven by cost savings and MediaTek’s strong relationship with TSMC, the world’s leading semiconductor manufacturer.

In 2024, Google introduced its sixth-generation TPU, positioning itself as an alternative to Nvidia’s powerful AI processors. However, despite being a competitor, Google still relies on Nvidia’s chips for some AI workloads. This partnership with MediaTek is part of its long-term strategy to reduce that dependence while strengthening its own AI hardware capabilities.

Research firm Omdia estimates Google spent between $6 billion and $9 billion on TPUs last year, based on Broadcom’s AI semiconductor revenue targets. This investment underscores Google’s push to establish itself as a key player in AI chip development, particularly as demand for Nvidia chips has surged among competitors.

But Google isn’t the only one moving toward independence. Meta is already testing its first in-house AI training chip, aiming to reduce its dependence on Nvidia. Meanwhile, OpenAI plans to finalize its first custom AI chip design this year, with hopes of mass-producing it at TSMC by 2026.

That being said, it will be interesting to see how this Google-Mediatek collaboration evolves and whether it will challenge Nvidia's dominance in the AI chip industry.

Meta is developing an in-house AI chip to train its AI model. Because, why not?

It could help it reduce reliance on other chip companies like Nvidia.

TechloyLouis Eriakha