Apple Intelligence Is Taking More Than 30GB on Some Macs

After you install macOS 27 Golden Gate, your Mac automatically downloads Apple’s on-device AI models the next time it connects to the internet, and unlike earlier macOS versions, there is no longer a way to stop it. Apple Intelligence has been growing steadily since iOS 26, but with macOS 27 the local models are effectively part of the operating system.

Apple’s own figures are up to 14GB for Macs with an M3 chip or newer and at least 12GB of unified memory, and up to 8GB for other compatible Macs. Real-world readings in System Settings run considerably higher, as MacRumors found on an M4 Pro Mac mini.

The M1 MacBook Air and MacBook Neo are particularly odd, since both sit in the category Apple lists at up to 8GB. Apple’s support document doesn’t explain whether its “up to 14GB” figure is directly comparable with the Apple Intelligence category in Storage settings, so 30GB is a reported extreme rather than what every Mac will see.

The models are needed because Apple Intelligence isn’t purely cloud-based. macOS 27 ships with the AFM 3 Core on-device model, while Macs with an M3 or newer and at least 12GB of memory can also use AFM 3 Core Advanced. Heavier requests can still be handed off to Private Cloud Compute.

You can still switch the feature off under System Settings → Apple Intelligence & Siri, but that doesn’t necessarily hand the space back right away, since macOS manages the models as system assets. Deleting files from system folders by hand is a bad idea. To check your own Mac, go to System Settings → General → Storage and click the info button beside macOS. A large System Data category alone doesn’t point to Apple Intelligence, as macOS lumps many resources together there. iPhones and iPads running iOS 27 and iPadOS 27 haven’t shown the same ballooning so far.

M Adnan Akram

M Adnan Akram is the founder and lead writer at Phoonomo, covering mobile technology with a primary focus on Apple and the broader smartphone industry, alongside emerging developments in AI. He follows product launches, spec leaks, and industry trends closely to keep readers informed with timely updates.