The Gemini-powered Siri might not use Apple’s servers after all

Published on:

Apple and Google are telling different stories about where the new Gemini-powered Siri will actually run. During Alphabet’s earnings call this week, CEO Sundar Pichai called Google Apple’s “preferred cloud provider” for developing Apple Foundation Models. That directly contradicts what Apple’s been saying about using its own Private Cloud Compute infrastructure.

Tim Cook said last week that Gemini Siri servers would run on Apple devices and Private Cloud Compute while maintaining Apple’s privacy standards. But Pichai’s statement suggests Google’s infrastructure will handle at least some of the processing.

Bloomberg reported shortly after the partnership announcement that Apple and Google were discussing hosting the chatbot directly on Google servers using TPUs (tensor processing units). Apple’s more immediate Siri update would stay on Private Cloud Compute, but the advanced Gemini version might take a different route.

What this means for privacy

Apple built its entire brand around keeping user data private. The company’s Private Cloud Compute system handles AI tasks that exceed on-device processing while maintaining privacy protections. Data processed through PCC isn’t stored, isn’t used for training, and isn’t accessible to Apple employees.

If Gemini-powered Siri runs on Google’s servers instead, that privacy promise gets complicated. Google would theoretically have access to user requests, even if both companies claim otherwise. Neither has directly addressed where processing actually happens.

The confusion might come down to timing. There could be two different Siri upgrades. The first update, expected soon, would run on Apple’s infrastructure. The more powerful “chatbot-style” Siri rumored for later this year might be the one moving to Google’s cloud.

Apple’s paying Google roughly $1 billion per year for this partnership. That’s on top of the $20 billion Google already pays Apple to be the default search engine. Analyst Ming-Chi Kuo says the deal is just buying time while Apple builds its own AI infrastructure.

Apple announced the partnership with vague language about maintaining privacy standards, but avoided specifics about server infrastructure. Until one of these companies clarifies what’s happening, users are left guessing about where their Siri requests will end up.

Source link

Related