Google as Apple Cloud

Google has publicly named Apple as a cloud customer and said Gemini will help power a more personalized Siri later in 2026. (blog.google) Thomas Kurian, chief executive of Google Cloud, repeated the partnership at Google Cloud Next on April 22 in Las Vegas after Google and Apple first announced a joint statement on January 12. Google said Apple is using Gemini models and cloud technology for the “next generation of Apple Foundation Models.” (blog.google) Apple’s “foundation models” are the core artificial intelligence systems behind Apple Intelligence, the company’s suite of writing, image, and assistant features. Apple’s developer documentation says those models also power the on-device model exposed through its Foundation Models framework. (developer.apple.com) That split matters because Apple has spent the past year selling Apple Intelligence as a mix of on-device processing and Private Cloud Compute, its own server system for heavier requests. The Google deal shows Apple is also leaning on an outside cloud provider for at least part of the next model stack. (apple.com; blog.google) Google has been telegraphing the relationship for months. In Alphabet’s fourth-quarter 2025 earnings remarks, Sundar Pichai said Google was collaborating with Apple as its “preferred Cloud provider” and on Apple Foundation Models based on Gemini. (blog.google) The immediate product target is Siri, which Apple had already said would become more personal by using a customer’s apps, messages, and on-screen context. Google’s January statement said those features were coming this year, and reports from April 22 said Kurian tied that same timeline to 2026 releases. (blog.google; macrumors.com) For Google Cloud, the Apple tie-in lands as the company pitches itself as a full artificial intelligence stack, from chips and servers to Gemini models and developer tools. Sundar Pichai said on April 22 that Google’s first-party models now process more than 16 billion tokens per minute through direct application programming interface use. (blog.google) For Apple, the arrangement gives it more computing capacity for large model training and inference without waiting for all of that infrastructure to be built in-house. Apple has not publicly detailed which Siri requests will stay on device, which will run on Apple servers, and which could touch Google Cloud-backed systems. (blog.google; apple.com) Neither company has published a technical architecture diagram or pricing terms, and Apple has not yet put the Google partnership in a standalone newsroom release. The next concrete checkpoint is Apple’s Worldwide Developers Conference in June, where Apple typically previews the next version of Siri and its operating systems. (apple.com; macrumors.com)