Apple is weighing a major change in how Siri works behind the scenes. The company is considering running its next-generation Siri chatbot on Google’s cloud instead of relying only on its own Private Cloud Compute servers. If this happens, it would mark a clear shift in Apple’s long-standing approach to privacy and infrastructure, driven by the growing demands of modern AI models.
According to Bloomberg’s Mark Gurman, Apple is in talks with Google to host the future Siri chatbot on Google-owned servers powered by Tensor Processing Units, or TPUs. These chips are built for large-scale AI workloads and can handle far more demanding models than Apple’s current setup. This discussion surfaced alongside reports about Apple’s plan to turn Siri into a full chatbot in iOS 27.
Apple is Considering Google’s Cloud
Apple’s existing Private Cloud Compute system was designed to extend on-device privacy into the cloud. It uses Apple-designed servers built around high-end Mac chips. Apple has said user data gets processed temporarily and is not stored or accessible, even to Apple. That promise has played a central role in Apple Intelligence messaging.
However, the upcoming Siri chatbot appears to need more computing power than Apple can easily provide today. The advanced version planned for the iOS 27 cycle is expected to rely on a newer large language model linked to Google’s latest Gemini technology. Running this model at scale strains Apple’s current cloud capacity, which explains the interest in Google’s much larger AI infrastructure.
What Changes and What Stays the Same
Apple is not abandoning its own cloud overnight. The near-term Siri upgrades in iOS 26.4 will still run on Apple’s Private Cloud Compute servers. These early features use less demanding models and fit within Apple’s existing system.
Here are the key differences being discussed:
- iOS 26.4 Siri updates run on Apple’s Private Cloud Compute
- iOS 27 Siri chatbot may run on Google servers with TPUs
- Newer models require far more processing power
- Apple would still control encryption and data handling
Privacy and Control Questions
Using Google’s cloud does not automatically mean Google would gain access to Siri conversations. Apple already works with third-party cloud providers for parts of iCloud while keeping control over encryption keys. In those arrangements, Apple decides how data gets handled and stored.
Apple would likely structure any deal with Google so that user data is not logged or retained. The goal would be to keep Siri requests isolated from Google’s advertising and data collection systems, even while using Google’s hardware.
A Practical Shift Inside Apple
This possible move reflects a broader change in Apple’s AI strategy. When Apple introduced Private Cloud Compute at WWDC 2024, it did not plan to rely on Google’s most advanced models. Since then, expectations around chatbots have shifted fast, driven by services like ChatGPT.
With new leadership overseeing Siri, Apple appears to be prioritizing speed and capability over strict adherence to its earlier vision. The pressure to deliver a competitive Siri experience is high. Using Google’s cloud may be the fastest way to close the gap, even if it means rethinking some core assumptions about how Apple runs its AI services.
