The latest update to Google Chrome introduces stronger on-device AI features that require significantly more disk space , up to 22 GB , posing challenges for users with smaller SSDs and highlighting a shift towards local processing at the expense of hardware flexibility.
Google Chrome’s built-in artificial intelligence is now imposing a larger storage burden on some users’ machines, with reports saying the browser may require as much as 20 GB of free space before its on-device features will install or run reliably. Windows Report said the updated guidance appeared on 1 August, while Google’s own developer documentation says Chrome needs at least 22 GB of free space on the volume that contains the profile data for its built-in AI features to work properly. The discrepancy appears to reflect different ways of describing the same storage headroom, but the practical point is clear: the browser is reserving a substantial amount of disk space for local AI functions.
The core model involved is Gemini Nano, a compact on-device system that can take up about 4 GB by itself, according to TechSpot and Computerworld. Chrome uses it for features such as writing help, page summaries, scam detection and tab organisation, with the processing done locally rather than sent to Google’s servers. That design reduces reliance on cloud processing and can improve privacy, but it also means the browser must check for sufficient free space, a stable internet connection and compatible hardware before it downloads the model.
For users with smaller SSDs, the storage requirement may be significant. TechSpot reported that the model can appear in Chrome’s directory and will be downloaded again if removed while on-device AI remains enabled. Google’s support pages say users who do not want the feature can turn off “On-device AI” in the browser’s System settings, which prevents Chrome from restoring the model. Google also advises against trying to delete the files manually, since the browser may simply fetch them again.
In managed environments, Chrome Enterprise administrators can control the feature through policy settings, giving organisations a more formal way to block or permit local AI use across fleets of devices. The broader picture is that Google is continuing to push more processing onto the device itself, but the trade-off is a larger storage footprint and tighter hardware requirements. For users working on laptops with limited disk space, that may matter as much as the convenience of the AI tools themselves.
Disclaimer: This content is intended for informational purposes only. Readers are advised to exercise their own judgement, conduct due diligence, or consult a qualified expert before acting on any information provided.
