Apple might spend upwards of $5 billion on servers alone within the subsequent two years because it makes an attempt to meet up with opponents within the generative AI area.
On Sunday, Bloomberg’s Mark Gurman reported that Apple will spend within the space of a billion {dollars} yearly on its Synthetic Intelligence efforts. Nevertheless, in a Medium put up on Monday, TF Securities tech business analyst Ming-Chu Kuo says the quantity Apple will spend could possibly be a lot greater.
In his put up, Kuo says a survey signifies that Apple has plans to buy 2,000–3,000 items of AI servers in 2023 and 18,000–20,000 items in 2024. He says Apple will account for about 1.3% and 5% of worldwide AI server shipments in 2023 and 2024, respectively.
Kuo says it’s “cheap to imagine” that Apple’s purchases of AI servers this 12 months, are primarily the Nvidia’s HGX H100 8-GPU, which is designed for generative AI coaching and inference. Kuo expects to see Apple improve its AI server purchases in 4Q24 to a “B100 answer.”
HGX H100 8-GPU servers go for a worth of round $250,000 apiece, that means Apple will spend roughly $620 million in 2023, and $4.75 billion in 2024 on AI servers alone — or about $5.37 billion within the subsequent two years.
Kuo says Apple would doubtless spend extra throughout 2023 if it weren’t for the present scarcity of Nvidia AI chips and Apple’s “putting its orders later than different main prospects.”
Whereas the quantity Apple will doubtless spend seems like an enormous outlay, Kuo believes Apple’s expenditures will lag behind opponents, as Meta is predicted to purchase round 40,000 AI server items in 2024, and Microsoft has plans to buy 80,000 to 100,000 items in the identical calendar 12 months.
Kuo says it’s “not applicable to match” Apple’s decrease variety of server purchases to Microsoft’s degree of purchases since it’s doubtless that Apple received’t supply companies like cloud internet hosting of AI tasks.
Kuo says he believes that when Apple trains its giant language fashions, it prefers to make use of bodily AI servers it has bought and put in itself somewhat than digital hosts from different cloud service suppliers. Kuo says that’s resulting from Apple’s want for safety, privateness, and design flexibility.
Whereas Apple’s AI infrastructure won’t present as a lot computing energy as its opponents, Kuo believes Apple thinks that it has “superior software program growth capabilities to catch up.”
Along with the value of the servers, there are a number of different bills that Apple might incur, akin to labor and infrastructure working prices. Kuo says it’s cheap to estimate that Apple will spend billions on an annual foundation to catch as much as the top begin its rivals at the moment get pleasure from.
Kuo additionally talked about Sunday’s report, saying, “If Apple actually intends to spend solely a billion {dollars} a 12 months on generative AI growth, it received’t matter a lot if my survey is fallacious, however I’m genuinely involved about the way forward for Apple’s generative AI enterprise/service.”
Kuo additionally sees Apple finally creating its personal server chips to avoid wasting on prices. Apple already makes its personal chips for many of its units, together with these utilized in its iPhone, iPad, Apple Watch, and Mac lineups. Nevertheless, Kuo mentioned, “The visibility of the event shouldn’t be clear at current.”