Overview
- Apple unveiled the new Siri AI at WWDC on June 8, released developer betas the same day, said a public beta will arrive next month, and tied a wider consumer launch to an autumn software update and new hardware.
- The architecture uses a system orchestrator to route requests between on‑device models and a Private Cloud Compute layer so routine tasks run locally while heavy inference runs in the cloud.
- Apple confirmed partnerships with Google Cloud and Nvidia to run its top-tier Apple Foundation Model Cloud Pro on Google’s infrastructure and Nvidia Blackwell GPUs, a deal that several outlets say may carry a large licensing cost though Apple has not confirmed a figure.
- Apple stresses privacy by keeping personal data processing on users’ devices and using hardware-backed confidential compute when outsourcing inference, but it will limit some features by device capability, language and region at launch.
- Analysts offered mixed readings: some see consumer upside and potential device upgrade demand, while others call the rollout cautious and note short-term investor concern after the WWDC announcements.