Market Analysis

Image Credit: Reuters

Apple has opted to use chips designed by Google (NASDAQ: GOOGL) rather than Nvidia (NASDAQ: NVDA), the leading provider of AI processors, for building key components of its upcoming suite of AI tools and features, according to a research paper published by Apple (NASDAQ: AAPL) on Monday.

This decision is significant given Nvidia's dominant position in the AI processor market, commanding approximately 80% share when including chips from companies like Google and Amazon.com (NASDAQ: AMZN). Although Apple’s paper did not explicitly state that it avoided using Nvidia chips, it did not reference Nvidia hardware in its description of the AI infrastructure.

Apple did not provide a comment on this development.

For training its AI models, Apple utilized Google's tensor processing units (TPUs) in large chip clusters. Specifically, Apple deployed 2,048 TPUv5p chips for its AI model designed for iPhones and other devices, and 8,192 TPUv4 processors for server-side AI models.

While Nvidia specializes in graphics processing units (GPUs) used extensively in AI, it does not design TPUs. Unlike Nvidia’s standalone chip sales, Google offers TPUs through its Google Cloud Platform, requiring customers to develop software via its cloud services to access the hardware.

This week, Apple began rolling out parts of its Apple Intelligence to beta users.

The extent of Apple's reliance on Google's hardware was fully revealed in the research paper, following an earlier report by Reuters in June.

Neither Google nor Nvidia commented on the matter. Apple’s engineers suggested in their paper that Google's chips could support even more advanced AI models beyond those discussed.

Apple had introduced several new AI features at its June developer conference, including the integration of OpenAI's ChatGPT technology into its software.

In regular trading on Monday, Apple’s stock declined by 0.1% to $218.24.

Paraphrasing text from "Reuters" all rights reserved by the original author.