Apple gains access to Google’s Gemini AI to power future Siri features

Apple has reportedly gained “complete access” to Google’s Gemini AI model, according to a new report from 9to5Mac citing The Information.

The technology is said to be deployed directly inside Apple’s own data centres. Developers are using a distillation approach, where the larger Gemini model transfers knowledge to smaller, more efficient models.

This allows Apple to build compact AI systems capable of running locally on iPhones while requiring significantly less computing power. Instead of simply copying responses, these smaller models learn to replicate Gemini’s internal reasoning processes.

At the same time, Apple continues developing its own foundation models, as Siri’s goals do not always align with Gemini’s strengths. The company is reportedly working on a major Siri upgrade that will include memory of past conversations and more proactive features.

The announcement is expected at WWDC in June, although some features may be delayed due to testing challenges. A full rollout could shift to May or September 2026.