Apple used its Worldwide Developers Conference event in Cupertino on June 8 to unveil a redesigned Siri and say some Apple Intelligence features run on its cloud model, AFM Cloud Pro, using Nvidia graphics processing units in Google’s cloud.

Executives said Apple’s system routes requests between on-device and cloud models based on computing needs and access to personal data.

This aligns with Apple’s position that Apple Intelligence defaults to on-device processing and uses cloud models only for more complex tasks.

Apple has said Private Cloud Compute encrypts requests directly from devices to validated server nodes and is designed for stateless processing, so personal data is not accessible even to Apple.

The announcement follows a multi-year collaboration Apple and Google entered in January on Gemini models and cloud technology to support future Apple Intelligence features, including a more personalized Siri later in 2026.

Google Cloud also said it keeps data encrypted while it is being processed, including for some AI workloads.

🔗 Source: CNBC

Recent Apple developments