Apple has announced a significant partnership with OpenAI to integrate powerful artificial intelligence (AI) capabilities into its voice assistant, Siri. However, a closer examination of a technical document released by Apple reveals that the company has been leveraging Google’s advanced AI hardware to train its own AI models.
To develop its foundational AI models, Apple engineers utilized their proprietary software framework alongside a variety of hardware, including Apple’s own on-premises graphics processing units (GPUs) and Google’s cloud-based tensor processing units (TPUs).
Google has been developing TPUs for approximately a decade, and its fifth-generation chips are renowned for their high performance in AI training tasks, rivaling Nvidia’s H100 AI chips. At its annual developer conference, Google announced that a sixth generation of TPUs is slated for release this year.
These TPUs are specifically designed to run AI applications and train complex models. Google has built a comprehensive cloud computing hardware and software platform around them, offering clients access to this technology through its cloud services, similar to how Amazon Web Services (AWS) and Microsoft Azure operate.
Apple and Google did not immediately respond to requests for comment regarding the extent of Apple’s reliance on Google’s AI hardware compared to other vendors like Nvidia. The use of Google’s chips typically involves purchasing access through Google’s cloud division.
Apple’s collaboration with Google underscores the interconnected nature of the tech industry’s advancements in AI. As companies strive to enhance their AI capabilities, partnerships and the utilization of shared resources become increasingly pivotal. This development signals a significant step in Apple’s efforts to catch up in the rapidly evolving field of artificial intelligence.
Reference(s):
cgtn.com
