Apple’s M2 Ultra AI Servers Could Be Ready for iOS 18 This Fall

Generative AI chip concept Credit: Unsplash
Text Size
- +

Toggle Dark Mode

In recent weeks, reports have circulated that Apple is building its own server farm to handle generative AI features that will be beyond the capabilities of on-device processing. Although the timeline for these hasn’t been entirely clear, a new report says that Apple hopes to have them ready in time for the launch of iOS 18 this fall.

News of Apple’s goals for AI servers first appeared last fall, when supply chain analyst Ming-Chi Kuo predicted that Apple could spend up to $5 billion to purchase several thousand units of AI servers. Kuo expected Apple to purchase off-the-shelf servers powered by Nvidia’s HGX H100 8-GPU for generative AI training.

This Limited-Time Microsoft Office Deal Gets You Lifetime Access for Just $39

Sick and tired of subscriptions? Get a lifetime license for Microsoft Office Home and Business 2021 at a great price!

However, recent revelations have turned that notion on its head as new information shows that Apple plans to use its own silicon instead of relying on Nvidia AI chips, which are already hard to come by.

Following news from analyst Jeff Pu that Apple has already tasked Foxconn with building AI servers powered by the M2 Ultra — the highest-performance Apple silicon currently available — the Wall Street Journal (Apple News+) reported yesterday that Apple has an internal project code-named ACDC — Apple Chips in Data Center — that’s been “in the works for several years.”

While the WSJ implied this was an entirely new chip, and couldn’t say when it would be revealed, Bloomberg’s Mark Gurman chimed in today, confirming Pu’s report that Apple will start with the stock M2 Ultra, and will likely move to an M4-based version as soon as practical.

WWDC2023 M2 Ultra

The first AI server chips will be the M2 Ultra, which was launched last year as part of the Mac Pro and Mac Studio computers, though the company is already eyeing future versions based on the M4 chip.Mark Gurman

Gurman adds that Apple’s plan to use its own chips to process AI tasks in the cloud was “hatched about three years ago,” but ramped up after ChatGPT spurred the recent AI craze.

However, Nvidia may not have been entirely kicked to the curb. As the WSJ notes, Apple’s in-house ACDC project is expected to be more about running AI models than training on them. Nvidia continues to dominate in that area, so Kuo’s information about Apple’s plans may be at least partly accurate.

Apple plans to continue doing as much on-device AI processing as possible, but complex generative AI could be beyond what even this year’s A18 Pro chip can muster.

Relatively simple AI tasks — like providing users a summary of their missed iPhone notifications or incoming text messages — could be handled by the chips inside of Apple devices. More complicated jobs, such as generating images or summarizing lengthy news articles and creating long-form responses in emails, would likely require the cloud-based approach — as would an upgraded version of Apple’s Siri voice assistant.Mark Gurman

In addition to Apple’s instinct to control as much of its processes as possible, the ACDC project also promises to protect user privacy in ways that outsourcing to third parties may not. For example, Gurman cites the Secure Enclave technology that’s used in the iPhone to protect payment cards and Face ID information as a method that Apple would use to protect data against potential security breaches.

Artificial Intelligence

The good news is that Apple isn’t planning on keeping folks waiting, as Gurman believes that Apple will announce these features as part of the iOS 18 rollout this fall. It’s uncertain whether we’ll hear about them at next month’s Worldwide Developers Conference (WWDC), as the back-end servers may not be ready for them to be operational during the iOS 18 beta cycle that will kick off in June.

However, Gurman also adds that some of Apple’s new AI features will only work on its latest devices, suggesting that the upcoming A18 chip and the M4 that launched in this week’s iPad Pro include “include significant upgrades to the so-called neural engine,” that will be required to drive the new AI capabilities it plans to introduced in its next major software updates.

[The information provided in this article has NOT been confirmed by Apple and may be speculation. Provided details may not be factual. Take all rumors, tech or otherwise, with a grain of salt.]

Sponsored
Social Sharing