Google Cloud CEO Thomas Kurian speaks at the Cloud Next ’26 keynote.
Credit: Google
On Wednesday, Google used its Cloud Next ’26 keynote to provide a clearer picture of how Gemini will power Apple’s revamped Siri later this year. The company said Apple is using Gemini-based models and Google Cloud as its preferred partner to build “Apple Foundation Models,” which will sit behind Apple Intelligence and the new, more personalized Siri.
On stage, Google Cloud CEO Thomas Kurian spoke in front of an Apple logo and said the two companies are working together to build next‑generation models that run across Apple devices and Apple’s own cloud systems.
Kurian said the work will help Siri become more conversational and context‑aware, using data on the device and in Apple’s Private Cloud Compute, while Google Cloud handles more complex training and inference tasks in the background.
Apple continues to stress its privacy controls and says it will create a strong separation between personal data and Google’s infrastructure.
Reports suggest that the upcoming Siri update—which has been delayed several times—will include deeper app actions and the ability to handle multi-step commands while maintaining what it already has: a chat-style interface. This update may also introduce a dedicated Siri app that keeps a record of conversation history, similar to Gemini.
Apple has not confirmed those interface details, but multiple outlets now describe them similarly, framing the project as a multi-year effort that could cost Apple around $1 billion per year.
For Google, this partnership means Gemini will power AI not only on Android devices but also on hundreds of millions of iPhones.
