Ahead of introducing AI features during its annual developers meet, Worldwide Developers Conference (WWDC) this June, Apple has reportedly begun working on its AI servers that use M2 Ultra chips, according to Jeff Pu, analyst at Hong Kong-based investment firm Haitong International Securities.
While rumors hint that Apple’s most generative AI features will run locally on the device, some of the AI offerings will run on the cloud, which is why it’s developing its own AI servers.
The report adds that Foxconn is playing a key role in putting together Apple AI servers featuring an M2 Ultra chip, with plans to enhance the server with an M4 chip by late 2025. Take note that we might see a glimpse of M4’s strength if the rumor about Apple employing an M4 chip in iPad Pro comes true, which could be released during Apple’s Let Loose event.
In addition, another rumor coming out of China on Weibo suggests that Apple is developing its own processors for AI servers, using TSMC’s 3nm chip process. The new chips might start mass producing by the second half of 2025, which aligns with the timeline when Apple is reported to utilize M4-powered servers. With this, Apple could leverage its own chips to boost the performance of its data centers and improve the efficiency of future AI applications based on the cloud.
Apple is one of the (and probably only) tech majors that has been late to the AI party, while its counterparts, Google and Samsung, have heavily optimized AI for smartphones. For instance, consider the Samsung S24 series or Pixel 8 series; these phones have made AI a backbone of their functionality, while Apple is still gearing up its features. We will know the rest of the story as in what Apple has been readying this WWDC.
Apple is not late to the AI party.
Apple is late to the AI tech reporting hypefest.
There is no AI party.