Search

Cookies

We use cookies to improve your experience. By continuing, you accept our use of cookies.

Technology

Google Partners Marvell on New AI Chips to Boost Efficiency & Rival Nvidia

· · 2 min read

Google and Marvell are reportedly collaborating on two new AI chips aimed at improving the efficiency of AI model operations. One chip may be a memory processing unit, while the other is a new Tensor Processing Unit (TPU).

Google's Strategic Move in AI Hardware

Google is reportedly partnering with Marvell Technology to develop two new artificial intelligence (AI) chips. This collaboration aims to significantly enhance the efficiency of running complex AI models, marking a crucial step in Google's ongoing efforts to solidify its position in the competitive AI hardware landscape.

New Chips Target Enhanced AI Performance

According to reports, one of the chips under development is expected to be a memory processing unit designed to complement Google's existing Tensor Processing Units (TPUs). The second chip is envisioned as a new, specialized TPU tailored specifically for AI model operations, promising greater power and speed for demanding AI tasks.

These advancements are crucial for Google as it seeks to strengthen its AI infrastructure and demonstrate tangible returns on its substantial AI investments, particularly within its growing cloud revenue segment. The company's strategy is to offer its TPUs as a compelling alternative to other leading AI accelerators.

Challenging Nvidia's Dominance

This initiative positions Google's TPUs as a direct alternative to Nvidia's widely adopted Graphics Processing Units (GPUs) in the AI market. The increased competition from a major player like Google could reshape the AI chip landscape and challenge Nvidia's long-standing lead in high-performance computing, fostering innovation across the industry.

Broader Semiconductor Ecosystem

Google's partnership with Marvell is part of a broader strategy to bolster its semiconductor ecosystem. The tech giant has been actively expanding collaborations with companies such as Intel and Broadcom to improve chip design and manufacturing capabilities, meeting the escalating demands of AI-driven infrastructure.

The design for the memory processing unit is anticipated to be finalized by next year, moving into test production shortly thereafter. Google's upcoming first-quarter earnings report, expected on April 29, will likely offer further insights into its AI investment strategy and competitive stance in the semiconductor space.

Related