Apple May Use Google Servers for AI Siri

This isn't just about licensing Google's AI models; Apple is reportedly discussing hosting the next generation of Siri on Google's own servers. This marks a major move for a company that has built its brand on controlling its own hardware and software stack. The potential deal underscores the immense infrastructural challenge of modern AI. Apple's own AI infrastructure, known as Private Cloud Compute, is reportedly only about 10% utilized on average. This suggests that even for Apple, building out the necessary server power to compete with Google's established AI-optimized infrastructure is a monumental task. Internal struggles have plagued Apple's AI development for years. Reports describe an AI/ML group dubbed "AIMLess" by employees and repeated changes in direction for Siri's development that have led to delays. This has forced Apple to look outwards to keep pace in the AI race. This potential server deal deepens an already complex relationship with Google. Apple already receives an estimated $8 to $12 billion annually from Google to make its search engine the default on Safari. While that deal is about search traffic, the new arrangement would embed Google's infrastructure into one of Apple's core services. Privacy remains a key negotiation point. Apple is said to be requiring that Google's servers meet its stringent privacy standards. Apple maintains that its "Private Cloud Compute" system offers users privacy guarantees, and any partnership would have to uphold these promises. For years, Apple has been a major client for public cloud providers, spending tens of millions a month on Amazon Web Services (AWS) and also using Google Cloud for services like iCloud storage. This potential Siri deal, however, represents a much deeper and more critical reliance on a direct competitor's infrastructure.