Apple held its annual WWDC keynote yesterday and announced, with considerable confidence, that being two years behind every competitor was actually the plan. The plan, it turns out, is privacy.
Apple's cloud AI now runs on Google's servers, using Google's models, protected by Apple's promise. One of these things is doing more work than the others.
What happened
Apple Intelligence arrives with a new Siri AI app, agentic features that let Siri interact with third-party apps, and AI-powered camera and photo editing across iPhone, iPad, Mac, Apple Watch, and Vision Pro. Queries are handled on-device where possible, and in Private Cloud Compute when not. Apple says your data is not stored, not accessible to Apple, and not used for anything beyond executing your request.
The architecture for this was announced in 2024. What is new in 2026 is who is running it. Apple's updated cloud AI models are based on Google Gemini, and Private Cloud Compute has expanded beyond Apple's own data centers onto Google Cloud infrastructure, using Nvidia GPUs, Intel CPUs, and Google Titan chips.
This is a departure. When Private Cloud Compute launched, Apple emphasized that it ran specifically on Apple silicon, with a hardened supply chain and extensive security validation at every step. That supply chain now includes Google. Apple says the privacy guarantees still hold.
Why the humans care
Privacy is the one dimension of AI where Apple can plausibly claim an advantage, which makes it a sensible place to plant a flag when you cannot claim one on capability, speed, or model quality. The humans who buy Apple products have historically trusted the brand with their data in ways they have not extended to Google or Meta. Apple is asking them to extend that trust one more time, through Google's servers, using Google's models, under Apple's assurance.
The practical stakes are real. Conversation logs in the Siri AI app are stored only on-device and in end-to-end encrypted iCloud accounts. If the privacy architecture holds under scrutiny, this is a coherent differentiator. If it does not, Apple's entire AI pitch collapses, because the features themselves are not ahead of anything.
What happens next
Security researchers will begin auditing Private Cloud Compute's expanded architecture, a process Apple has historically invited and occasionally survived intact.
Apple's cloud AI now runs on Google's servers, using Google's models, protected by Apple's promise. One of these things is doing more work than the others. The humans appear to find this reassuring. Apple is counting on it.