Updated
Updated · Computerworld · Oct 28
Opera Rolls Out WebAI Preview for Local Browser Models on 64-bit PCs and Apple Silicon Macs
Updated
Updated · Computerworld · Oct 28

Opera Rolls Out WebAI Preview for Local Browser Models on 64-bit PCs and Apple Silicon Macs

3 articles · Updated · Computerworld · Oct 28

Summary

  • Opera has started an opt-in early preview of WebAI in Opera One, letting websites and extensions use AI models running locally in the browser instead of sending data to cloud providers.
  • 64-bit Windows PCs with Intel or AMD chips and Apple silicon Macs are supported, while Windows on Arm is not; models download only after users enable the feature, with site-specific permissions controlling access.
  • Google Gemma and Apple’s on-device foundation model are among the supported models, and Opera plans more device-optimized options as it expands the technology.
  • For enterprises, local processing could cut cloud AI usage and reduce exposure of sensitive data, but analysts warned it may be harder for IT teams to monitor device-level activity and block malicious sites or rogue extensions.
  • Smaller models could handle tasks like summarization on existing PCs, potentially delaying some AI hardware upgrades, though more advanced workloads are still expected to rely on dedicated hardware or cloud systems.

Insights

Opera made local AI opt-in and site-specific, but is that enough to make browser-based models trustworthy for enterprise use?
If browsers start running AI locally, will companies save on cloud costs and delay hardware upgrades, or just inherit harder security problems?