Apple unveils AI architecture built around Google Gemini models

Apple today announced a major overhaul of Apple Intelligence, unveiling an architecture built on foundation models developed in collaboration with Google using the technologies behind the Gemini family of models. The company described the work as a deep collaboration rather than a simple licensing arrangement, and said it delivers what it calls a huge upgrade to its system-level AI features.

Models built with Google, run on-device and in the cloud

The new architecture centres on Apple Foundation Models co-developed with Google. According to Apple, the models are adapted to run both on-device and on servers through its existing Private Cloud Compute infrastructure, the privacy-oriented layer that bridges local processing and data-centre scale compute.

The upgraded models bring state-of-the-art understanding and reasoning, along with multimodal support that includes image understanding and generation. Apple highlighted realistic image creation, advanced photo editing and visual question answering as new use cases — areas where the company had lagged behind rival assistants.

A higher-power tier, but no device list

Certain devices will receive a higher-power version of the model with additional capabilities, including speech generation, improved dictation accuracy and stronger natural language understanding. Apple did not specify which devices qualify for that tier, leaving open the question of how far back supported hardware goes.

An orchestrator in the middle

A new system orchestrator sits at the centre of the revised architecture, coordinating Apple Intelligence features securely across Apple's platforms. Apple says the orchestrator lets the system tailor responses based on the active app and the user's current task, an approach it describes as truly system-wide intelligence.

Apple used the announcement to frame its approach as a contrast to competitors it characterised as racing forward without regard for users. The company reiterated that Apple Intelligence relies on on-device processing and Private Cloud Compute, with a promise that user data is used only to execute the immediate request and is not accessible to Apple or third parties. Apple added that outside experts can verify those privacy guarantees at any time.