Apple Explores AI Chip Acquisitions to Reduce Nvidia Dependence and Boost In-House AI Development

Apple, long reliant on Nvidia for advanced artificial intelligence (AI) chip capabilities, is reportedly considering acquiring AI chip companies to lessen this dependence and accelerate its own AI infrastructure development. Currently, Apple uses its own chips for some AI workloads in its data centers but depends on Nvidia hardware via Google Cloud for more complex tasks, including running Google's Gemini model that powers the latest Siri version.

Apple's internal server chip, codenamed Baltra, expected this year, has faced delays, prompting the company to explore external acquisitions. This marks a shift in Apple's acquisition strategy, which traditionally favored smaller deals; earlier this year, Apple completed a nearly $2 billion purchase of Israeli AI firm Q.ai, specializing in speech decoding through facial muscle movements, one of its largest acquisitions after Beats Electronics in 2014.

Financially, Apple signaled greater flexibility, with CFO Luca Maestri indicating the company will no longer strictly maintain a "Net Cash Neutral" policy, potentially freeing capital for significant acquisitions. Alongside external acquisitions, Apple continues internal development of server chips based on M5 Ultra and the future M7 Ultra, expected to double memory capacity but not ready before 2029.

The need for improved server capabilities became evident during new Siri development, as Apple's chips could not handle large Gemini models, necessitating Nvidia chips on Google's cloud. Apple is also collaborating with Broadcom on AI server chips, extending their partnership through 2031, and is exploring acquiring AI companies to develop smaller, efficient models for iPhones.

Leadership changes may drive a more aggressive acquisition approach, with John Ternus set to succeed Tim Cook as CEO in September and Johnny Srouji gaining broader hardware engineering responsibilities. Expanding AI chip acquisitions would signal Apple's intent to build an independent alternative to Nvidia in the coming years.