CUPERTINO / LONDON (IT BOLTWISE) – Apple has introduced a new feature with MacOS Tahoe 26.2 that makes it possible to connect multiple Macs into a powerful AI cluster. This innovation is aimed at developers and researchers who want to run large language models locally and require significant computing power to do so.
Today’s daily deals at Amazon! ˗ˋˏ$ˎˊ˗
With the introduction of MacOS Tahoe 26.2, Apple has taken a new path to meet the increasing storage and processing power requirements for AI applications. The new feature allows multiple Macs to be connected into a cluster via Thunderbolt 5, which is of particular interest to developers and researchers who want to run large language models locally. This development could fundamentally change the way AI development work is done on Mac hardware.
A notable aspect of this innovation is the ability to connect four Mac Studio devices into a cluster that can share up to 2 TB of unified memory. This is enough to run the Kimi-K2 thinking model with a trillion parameters. Compared to previous clustering attempts, which often failed due to bandwidth limitations, the new solution leverages the full 80 GB/s capacity of Thunderbolt 5.
The hardware requirements for this clustering feature are specific: It is compatible with M4 Pro Mac Mini, M4 Pro/Max Macbook Pro, and Mac Studio models with Thunderbolt 5. Users only need standard Thunderbolt 5 cables to connect the Macs, without the need for specialized networking hardware. Apple demonstrated a four Mac Studio cluster running the Kimi K2 Thinking model while consuming less than 500 watts.
Interestingly, the MacOS Tahoe 26.2 update also brings MLX framework support for the M5 chip’s Neural Accelerators. This integration could improve AI inference performance on compatible hardware, although testing with the M5 MacBook Pro is still pending. Apple’s open source project MLX is becoming increasingly important among developers working with machine learning on Mac hardware.
The costs of implementing such a cluster are not insignificant. A Mac Studio with 512 GB of unified memory and an M3 Ultra chip starts at just under 12,000 euros. Still, the investment could be worth it for certain applications, especially when considering the benefits of locally processing large language models.
Order an Amazon credit card without an annual fee with a credit limit of 2,000 euros!
Bestseller No. 1 ᵃ⤻ᶻ “KI Gadgets”
Bestseller No. 2 ᵃ⤻ᶻ “KI Gadgets”
Bestseller No. 3 ᵃ⤻ᶻ “KI Gadgets”
Bestseller No. 4 ᵃ⤻ᶻ “KI Gadgets”
Bestseller No. 5 ᵃ⤻ᶻ “KI Gadgets”
Please send any additions and information to the editorial team by email to de-info[at]it-boltwise.de. Since we cannot rule out AI hallucinations, which rarely occur with AI-generated news and content, we ask you to contact us via email and inform us in the event of false statements or misinformation. Please don’t forget to include the article headline in the email: “Apple turns Macs into powerful AI clusters”.