In a surprising shift, Apple’s latest research report reveals that the tech giant has turned to Google for training two pivotal AI models, sidelining Nvidia‘s industry-leading chips.
Released on Monday, Apple’s report highlights the company’s reliance on Google’s tensor processing units (TPUs) for its AI infrastructure. This marks a significant departure from the norm, as Nvidia, which commands approximately 80% of the market share with its graphics processing units (GPUs), has been the dominant player in the AI chip sector.
The report conspicuously avoids mentioning Nvidia’s hardware, sparking speculation about Apple’s strategic pivot. Instead, Apple detailed its use of Google’s TPUs, which are deployed in extensive chip clusters to power AI model training.
While Nvidia specializes in GPUs, which are widely adopted in AI applications, it does not manufacture TPUs. Google, in contrast, provides access to its TPUs exclusively through its Google Cloud Platform. This setup requires users to develop software within Google’s ecosystem to leverage the TPUs.
This move underscores Apple’s strategic realignment and raises questions about the future dynamics in the competitive landscape of AI hardware.
Related topics: