Apple May Use Google Servers for Siri

This potential partnership is part of a broader, multi-year deal reportedly worth around $1 billion annually for Google to license its Gemini AI models to power a more intelligent Siri. The move to also use Google's servers would represent a significant expansion of this collaboration. Historically, Apple has maintained a "frenemy" relationship with Google, most notably through a deal where Google pays an estimated $20 billion per year to be the default search engine in Safari. This existing financial tie highlights a long-standing collaboration between the two tech giants, despite their intense competition in the smartphone market. Apple's consideration of Google's servers stems from a need to quickly scale its AI capabilities. Reports suggest Apple's own internal AI infrastructure was "beginning to decay" and the chips in its "Private Cloud Compute" servers are not optimized for large AI models like Gemini. To address privacy concerns, Google would be required to set up specialized servers that adhere to Apple's stringent privacy standards. The proposed "Stateless AI" system would process Siri queries without storing the data on Google's side, aiming to maintain user privacy. This isn't Apple's first time relying on rivals for cloud infrastructure. The company already uses a hybrid model, spending over $30 million monthly on Amazon Web Services (AWS) and also utilizing Google Cloud for services like iCloud storage. Apple's AI strategy, internally codenamed "Project Greymatter," prioritizes on-device processing for speed and privacy on tasks it can handle locally. The move to potentially use Google's servers is for more complex AI queries that require the power and scale of cloud computing.