google-chrome ai-features storage-management llms

Chrome's AI features may be hogging 4GB of your computer storage

theverge.com

May 10, 2026

2 min read

🔥🔥🔥🔥🔥

52/100

Summary

Google Chrome may automatically download a 4GB weights.bin file to its system folders when certain AI features are enabled, leading to unexpected drops in available storage for users. This on-device AI model file is installed in the browser directory, contributing to increased storage usage.

Key Takeaways

Google Chrome may automatically download a 4GB weights.bin file related to the Gemini Nano AI model when certain AI features are enabled, impacting users' available storage.
The Gemini Nano model is designed to run locally for privacy benefits, but users are not clearly notified about the storage requirements when enabling these features.
To free up space, users must disable the On-Device AI option in Chrome settings, as simply deleting the file will result in it being re-downloaded.
Google has stated that the Gemini Nano model will uninstall automatically if the device is low on resources and has provided a way for users to turn off and remove the model directly in Chrome settings.

Read original article

View story →

Coverage & discussions

Hacker News · theverge.com

Chrome's AI features may be hogging 4GB of your computer storage ↗

Community discussions

Hacker News discussion ↗

Historical count: 5 · 5 saved in sample · Historical sample

Analysis ready

Community Sentiment

Negative

Positives

The local model's presence suggests a potential for enhanced privacy, albeit at a significant storage cost, which could lead to more user control over data.
Switching to Safari has yielded positive experiences for some users, indicating that alternatives to Chrome may offer better performance and lower resource usage.

Concerns

The 4GB storage requirement for Chrome's AI features raises concerns about unnecessary resource consumption, suggesting a lack of efficiency in design.
Users express skepticism about Google's intentions regarding privacy, indicating a distrust in the company's commitment to user data protection.
The reliance on a local model for features that could be handled server-side implies that Google prioritizes its own compute savings over user experience.