Apple reveals new AI architecture built around Google Gemini models
Original: Apple reveals new AI architecture built around Google Gemini models
Why This Matters
Represents Apple's strategic partnership with Google to compete in AI capabilities while maintaining privacy focus
Apple announced a major overhaul of its Apple Intelligence platform featuring a new architecture built on foundation models co-developed with Google using Gemini technologies. The upgrade brings multimodal support, image generation, and advanced reasoning capabilities while maintaining on-device processing and Private Cloud Compute infrastructure.
Apple unveiled a significant transformation of Apple Intelligence, introducing foundation models developed in collaboration with Google based on Gemini family technologies. The new architecture supports both on-device and server processing through Apple's Private Cloud Compute infrastructure. Key enhancements include realistic image creation, advanced photo editing, visual question answering, and multimodal capabilities. Select devices will receive higher-power model versions with speech generation, improved dictation accuracy, and enhanced natural language understanding. A new system orchestrator coordinates features across Apple platforms, enabling system-wide intelligence tailored to active apps and user tasks. Apple emphasized its privacy-focused approach, contrasting with competitors it described as 'racing forward' without user consideration, reiterating that user data serves only immediate requests and remains inaccessible to Apple or third parties.