Apple
is gearing up for its own push into artificial-intelligence technology to match Google and
Microsoft.
The iPhone maker will have to radically step up investment to support the effort.
Apple
(ticker: AAPL) is likely to spend billions on AI chips from Nvidia (NVDA) to power that effort, according to Ming-Chi Kuo, an analyst at TF International Securities.
Apple will spend at least $620 million on AI servers this year, rising to $4.75 billion next year, according to Kuo. That’s based on estimated purchases of 2,000-3,000 AI server units this year and 18,000-20,000 units next year.
Apple didn’t immediately respond to a request for comment from Barron’s early on Tuesday. Apple shares were down 0.5% in early trading.
Kuo’s assumption is that Apple will be buying the most popular AI server specification, based on eight Nvidia H100 chips networked together. The current shortage of Nvidia chips means the major purchases will only take place in 2024, he wrote.
That would still leave Apple’s AI investment well below that of some of its rivals.
Facebook
-owner Meta Platforms (META) has said it has spent billions on AI infrastructure already and it will be its single largest investment going forward. Kuo estimated Meta will purchase around 40,000 AI server units in 2024.
Apple has generally shied away from public pronouncements on its AI goals. It doesn’t have its own business offering access to AI models via a cloud-computing platform like Microsoft (MSFT) or Google-parent
Alphabet
(GOOGL) and hasn’t publicly unveiled its own in-house large language model.
However, Bloomberg reported recently that Apple is planning to spend around $1 billion a year on developing generative AI, including a new version of Siri that could be ready next year, citing people familiar with the matter.
Some evidence of Apple’s plans could come with its fiscal fourth-quarter earnings report on Nov. 2.
Write to Adam Clark at adam.clark@barrons.com
Read the full article here