Apple could be set to spend around $1 billion on Nvidia GB300 NVL72 systems.
As reported by Investor’s Business Daily, Loop Capital analyst Ananda Baruah said that the company’s order represents around 250 servers sold at $3.7m to $4m each.
According to Baruah, the company will use servers from Dell Technologies and Super Micro Computer (SMCI) powered by the Nvidia GB300s for the cluster that will be used for generative AI applications.
“Apple is officially in the large server cluster Gen AI game … and SMCI & Dell are the key server partners,” he said. “While we are still gathering fuller context, this appears to have the potential to be a Gen AI LLM (large language model) cluster.”
Baruah speculated that the investment could be motivated by Apple’s struggle to bring its AI-enabled Siri assistant to market, which it had hoped would be available early this year.
DCD has reached out to Apple for comment.
A large-scale investment in Nvidia chips is a shift in Apple’s AI strategy thus far.
In July 2024, the company published a research paper that suggested it was using Google’s Tensor Processing Units (TPUs) to train its AI models, omitting any mention of Nvidia GPUs.
Later that year, in December, the company was said to be working with Broadcom to develop its own AI server chips dubbed Baltra. Reports suggested the Apple chips would be developed using TSMC’s N3P advanced manufacturing process, with an aim for mass production by 2026.
The Nvidia GB300 NVL72 is a fully liquid-cooled rack-scale offering. Its features 72 Nvidia Blackwell Ultra GPUs, and 36 Nvidia Grace CPUs. The Blackwell Ultra was revealed at Nvidia’s GTC event earlier this month. Nebius and YTL are set to offer the hardware via their cloud offerings, among others.
