Apple says its AI is still private, even when it's running on Google's servers
Ars Technica AI · Andrew Cunningham · 2026-06-09
Apple confirmed at WWDC that Siri AI uses Google Gemini models running on Nvidia hardware in Google's data centers, yet Apple continues to make the same privacy promises it made when AI processing ran on Apple-controlled infrastructure.
Appears in
Extraction
Topics: apple-intelligenceai-privacycloud-aisiri
Claims
- Apple confirmed at WWDC that Siri AI uses Google Gemini language models running on Nvidia hardware installed in Google-operated data centers.
- Apple is maintaining identical privacy promises for Siri AI despite processing shifting from Apple-controlled hardware to Google's third-party infrastructure.
- Apple's Private Cloud Compute system, which anchored prior privacy guarantees, relies on Apple's own server hardware and cannot scale to meet Siri AI's capacity requirements.
- On-device language models on iPhone and Mac are too small to power the reasoning capabilities Siri AI requires.
Key quotes
Apple says its AI is still private, even when it's running on Google's servers.
Apple has run up against the limits of its own hardware.
Apple would have had to commit to a huge data center buildout that it has so far avoided.