The reason why Google Chrome requires 20 GB of free space to run local AI models and how you can disable this feature is in our news.
Google has updated its official documentation regarding the hardware requirements for its AI integrations on Chrome. The company mandates at least 20 GB of free disk space on the system to ensure efficient use of on-device AI features. This includes the automatic download and installation process for local models like the Gemini Nano.
Google also considers criteria such as internet connection and hardware performance, and only enables AI components on computers that meet these requirements. Users now have more control over this process, which takes up valuable space on their disks.
Why are Disk Requirements Kept High?
According to statements made by Google, the basic data size required for AI models to run is approximately 4 GB. However, the reason the system requires a considerably high amount of free space, over 20 GB, is due to the additional storage space needed for post-installation updates, data processing, and the stable operation of the model.
The browser seeks ample space for these background processes.
On devices with insufficient disk space, the loading of AI models is automatically blocked.
Users can disable Native AI.
Many users may be bothered by these background data download processes in the browser. To prevent this, Google has added a special “On-device AI” switch to Chrome settings. When this feature is disabled, the browser deletes models that have already been downloaded and stops new downloads.
If the settings are changed, Chrome scans the hardware capacity again and downloads the necessary documents again to re-enable the system. This process can be managed more strictly for corporate users with Chrome Enterprise policies.
Google warns that manually deleting AI files can mislead the system.
System Requirements May Change in the Future
The tech giant states that current requirements are not fixed and may change with future updates. In particular, as AI models develop and their capabilities increase, the resources needed by the browser are expected to increase as well. While browser-based AI aims to optimize user experience, these new standards may be a concerning development for users of devices with limited storage space.
Do you think allocating 20 GB of disk space for browser-based AI models is a reasonable request, or should Google offer more optimized solutions in this regard? You can share your opinions with us in the comments section.