Unlike other major tech companies making significant capital investments in AI data centers, Apple has long avoided this approach. However, this seems to be changing. According to Investor’s Business Daily (IBD), Apple is now ramping up investments in AI infrastructure.
Loop Capital analyst Ananda Baruah reports that Apple has placed orders for approximately $1 billion worth of NVIDIA GB300 NVL72 systems. The order includes around 250 systems, each priced between $3.7 million and $4 million. Additionally, Apple is collaborating with Dell and Supermicro to develop its own large server cluster for generative AI applications. Although details remain scarce, it appears that the company aims to use this cluster for large language models (LLM). Previously, Apple partnered with Google Cloud, utilizing TPUs to train AI models, while relying on its M-series chip-powered servers for secure inference. The company is also rumored to be developing its own AI accelerators with Broadcom’s support.
Delays in AI Rollout and Leadership Changes
Baruah suggests that Apple’s shift in strategy stems from difficulties in launching its AI-powered Siri assistant, first announced at WWDC 2024. The new Siri features were initially expected to launch earlier this year but are now projected for release in 2026.
According to Bloomberg, an Apple executive called these delays “disgusting” and “shameful,” particularly since the company had actively promoted AI features in TV advertising. Apple is already facing lawsuits related to these issues, reminds NIX Solutions. In response, the company replaced John Giannandrea with Mike Rockwell, the head of Vision Pro development, as the new leader of its AI division.
Apple’s evolving AI strategy signals a major shift in its approach to artificial intelligence. We’ll keep you updated as more developments unfold.