Apple May Use Google Servers for AI Siri

This potential deal is part of Apple's broader hybrid strategy for "Apple Intelligence," which prioritizes running AI processes directly on devices for speed and privacy. For more complex requests, Apple uses a system called Private Cloud Compute, and this partnership with Google would handle the most demanding AI tasks that exceed its current infrastructure's capabilities. Talks between the two tech giants reportedly center on Apple licensing Google's family of AI models known as Gemini. A multi-year deal, potentially worth $1 billion annually, would have Apple use Google's technology to power a revamped Siri and other AI features, expected to roll out in 2026. This collaboration appears to extend beyond just licensing software. Apple is reportedly asking Google to host and operate dedicated server clusters for Siri inside Google's data centers. This signals a major strategic shift for Apple, which has historically focused on building and controlling its own server infrastructure to protect user privacy. To address privacy concerns, Google would be required to adhere to Apple's strict standards. One potential arrangement involves a "managed server" model, where Apple would operate its own customized hardware within Google's facilities, leveraging Google's infrastructure while maintaining control over data security. While leaning on a rival for critical infrastructure, Apple is also developing its own server-grade silicon. A secretive internal effort, "Project ACDC" (Apple Chips in Data Center), has been working for years to create custom chips specifically for AI processing in its data centers. This is not the first major partnership between these competitors. Google already pays Apple an estimated $20 billion a year to remain the default search engine in the Safari browser. This new AI deal would further entangle the two rivals, creating a deeper financial and technological dependency.