Apple is considering a major shift for its next-generation Siri chatbot. According to a Bloomberg report, the company is in talks with Google to host the advanced AI on Google's cloud servers. These servers would use Tensor Processing Units (TPUs), custom chips designed for large-scale AI workloads. This arrangement marks a departure from Apple's current strategy of processing requests on-device or through its own Private Cloud Compute infrastructure.

The immediate Siri updates in iOS 26.4 are still expected to run on Apple's own Private Cloud Compute servers, which use high-end Mac chips. However, the more advanced chatbot planned for iOS 27 may require Google's specialized cloud footprint and hardware. The new model, internally called Apple Foundation Models version 11, is comparable to Google's Gemini models and could exceed the capacity of Apple's current infrastructure.

Apple already uses third-party cloud providers like Google for parts of iCloud, retaining control over encryption keys. This potential partnership suggests Apple is prioritizing capability over keeping all infrastructure in-house. The move highlights the immense computational demands of modern large language models. The final decision will be watched closely by the tech industry for its implications on AI competition and cloud partnerships.