Apple is taking a cautious approach to generative AI, leveraging its existing M2 Ultra chips in data centers before transitioning to its next-generation M4 chips.
Apple believes that existing M series chips have sufficient security features
According to Bloomberg, Apple plans to offload complex AI queries to M2 Ultra processors running in their data centers. This strategy comes after an earlier plan, codenamed Project ACDC (Apple Chips in Data Center) which envisioned custom chips designed specifically for data centers. However, the trillion dollar company now believes its existing M series chips possess sufficient security features for their AI needs.
These M2 Ultra chips will be deployed initially in Apple’s own data centers, with the potential for expansion to third-party servers in the future. Apple has a network of data centers across the US, with a new facility currently under construction in Waukee, Iowa.
While competitors like Google, Meta, and Microsoft have aggressively pursued generative AI, Apple has focused on research and development. In December, the company’s machine learning team released MLX, a framework designed to optimize AI models for Apple silicon. They’ve also published research exploring potential applications of AI on devices and how it could enhance existing features like Siri.
The recent emphasis on AI performance during the M4 chip announcement, highlighting its “outrageously powerful” neural engine, suggests Apple is gearing up for a more prominent role in the generative AI landscape.
RELATED:
- Apple Pencil Pro vs Apple Pencil 2nd Gen: Don’t Be Fooled by Appearances
- What is Tandem OLED? Apple iPad Pro 2024’s New Ultra Retina XDR Display Explained
- Xiaomi 13 Ultra Premium Camera Phone is now only $799
- REDMAGIC Magnetic VC Cooler 5 Pro now available at a discounted price at GeekWills
- 6 Amazing One UI 6.1 AI Features You Can’t Ignore
(Via)