According to reports, Apple is greatly investing in Nvidia machines for AI development.

In a server, the Siri image appears.
An analyst claims the company is now spending$ 1 billion to purchase Nvidia systems as well, perhaps in opposition to Apple’s very public claim that it is using Apple Silicon servers to run Apple Intelligence. 13 / 24: There was a rumor that Apple had run its Artificial servers using its own Apple Silicon computers in April 2024. The assumption was therefore that all data centres would get run on Apple Silicon bits in June 2024. &# 13, By September 2024, it was certain. Craig Federighi of Apple claimed that Apple’s Intelligence machines were located on Apple Silicon, which was crucial for keeping the company’s AI services secret. &# 13, Yet now, according to Loop Capital analyst Ananda Baruah, as reported by Investor’s Business Daily, Apple is reportedly planning to spend about$ 1 billion on new Nvidia servers specifically for generative AI. In a note to investors, he wrote in &# 13, “AAPL is officially in the large server cluster Gen AI game and SMCI]Super Micro Computer ] & Dell are the key server partners.” This appears to have the potential to be a Gen AI LLM ( large language model ) cluster, according to Baruah, who claims that Apple is purchasing 250 Nvidia NVL72 servers for between$ 3.7 million and$ 4.2 million each. Nvidia claims that its NVL72 client contains 36 Grace CPUs and 72 Blackwell GPUs. Additionally, the business claims that this site is not yet accessible as of March 18, 2025. It’s no wonder that Apple sees the need to develop its machines, and it’s doubtless possible that it already has the machines pre-ordered. There is no way to show for the moment, but given the numbers, this might be for growth purposes, assuming the report is accurate. This just doesn’t quite fit with Federighi’s claim that using Apple Silicon servers” sets a new standard for processing in the cloud in the industry” ( 13 ), he said.” When we didn’t have any before, we were building a custom OS to run in the data center. One of the most important and crucial components of the solution was “[Creating ] the trust model where your unit won’t refuse to issue a plea to a site unless the name of all the program the client is running has been published to a accountability logs. 

About The Author

See author's posts