Google would like to strengthen its AI chips with a new supplier / CheapToken News

Published on April 20, 2026 at 11:35

Comment

This partnership would aim to diversify Google's semiconductor supply chain and compete with Nvidia in the inference field.

Google would be in talks with Marvell Technology to develop two new AI chips: a memory processing unit and an optimized TPU for inference. The American company would thus aim to diversify its supply chain for custom semiconductors. She would rely on Broadcom, MediaTek, and now Marvell for design, while entrusting manufacturing to TSMC.

A partnership between Google and Marvell for the design of two chips

At the moment, no contract has been signed, but according to the American media The Information, Google is in close talks with Marvell Technology to develop two new chips aimed at running AI models. The first chip is a memory processing unit optimized to assist Google's existing TPU chips. The second is a new generation of TPU dedicated to inference: the operational phase where the model, once trained, is used to generate predictions or responses in real-world conditions.

A third supplier for Google

This latter chip therefore aims to make Gemini faster and cheaper. Through this partnership, Google is focusing on inference as the main area of computational expense, and hopes to thus compete with Nvidia in this field. It is worth noting that Marvell would act as a design services provider for Google.

A third supplier for Google

By partnering with Marvell Technology, Google is not replacing Broadcom, but adding a design partner to its supply chain, which includes Broadcom (high-performance chips), MediaTek (notably the "e" series versions, optimized for better performance-to-price ratio), and TSMC (manufacturing). It is therefore about a strategy of diversification for Google.

Ironwood, the seventh-generation TPU of Google, designed by Broadcom, was launched this month. Google presents it as « the first Google TPU designed for the inference era ». The two chips developed by Marvell Technology will therefore complement those of Broadcom, without replacing them. The goal is simple: to allow Google to adjust its resources according to the types of tasks or budgets. This addresses the increasingly significant portion of computing dedicated to inference, i.e., model deployment, rather than their training phase.

Google has also signed an agreement with the artificial intelligence company Anthropic. This latter will have access to a gigawatt of AI computing capacity, thanks to Google's AI chips called « Tensor Processing Units », or TPU.