When Apple announced that it was going to bring a new Gemini-powered Siri soon, many were surprised. The Cupertino giant had, for the first time, decided to outsource AI technology instead of continuing to work on its own projects for Apple Intelligence. And now, it seems that Apple is willing to rely more on Google for Siri.
As per The Information, Apple has asked Google to investigate running Gemini-powered Siri on its Google cloud infrastructure. However, instead of existing servers, Google may set up new servers that meet the Cupertino giant’s privacy requirements.
Why is Apple considering Google servers for Siri?
Apple is set for a major overhaul to Siri. The tech giant will reportedly pay $1 billion per year to Google to use its Gemini AI models for Siri. This collaboration is expected to bring new features to Apple’s personal assistant, which has so far lagged behind when compared to rivals like Gemini and newer players like ChatGPT.
While the Cupertino giant mentioned that Siri will run on the company’s private cloud computer, the report indicates that Apple is struggling with its own cloud infrastructure.
Apple’s private cloud compute is based on Macs. These Macs are likely not as powerful as Google’s cloud servers, that are run on Tensor Processing Units (TPUs). TPUs are Google’s own hardware, specifically designed for running Large Language Models (LLMs) like Gemini.
This could be the potential reason why Apple is considering having Google run Siri. The AI assistant will be powered by Gemini, and Google’s TPUs could likely not only be more powerful but also more efficient.
Apple’s AI infrastructure is starting to decay
As per the report, Apple’s own AI infrastructure is “beginning to decay.” The Cupertino giant is believed to have realised that it needed external support for cloud compute during its process of decommissioning old Nvidia-powered servers.
Apple is said to be using only around 10 per cent of its total capacity. It is believed some servers for private cloud compute are still stored in warehouses. The Cupertino giant has turned to Amazon and Google for cloud computing in the past.
In this scenario, using Google’s compute could allow Apple to meet the expected surge in Siri demand following its overhaul without incurring massive capital costs – similar to the cost savings it obtained by using Gemini to power Siri instead of continuing to develop its own models.
When will Apple launch Gemini-powered Siri?
Apple is expected to bring a major update to Siri with iOS 26.4 soon. This update will bring more features to the personal assistant using Google’s foundational models. However, the Cupertino giant is expected to transform Siri into a ChatGPT-like chatbot with iOS 27 that will likely be launched in September this year, alongside the iPhone 18 series release.
