Apple and Google have agreed a multi-year collaboration that will base the next generation of Apple Foundation Models on Google’s Gemini models and cloud technology.

The companies said the work will feed into future Apple Intelligence features. They also said it will include a more personalised Siri that the companies said will arrive this year.

Apple Foundation Models sit at the centre of Apple’s generative AI strategy and are used across system-level experiences under the Apple Intelligence banner. Google’s Gemini is the company’s family of AI models, which it offers for consumer and business use and through its cloud platform.

The partnership brings two of the world’s largest consumer technology companies into a closer operational relationship on a strategically sensitive layer of the software stack. It also underlines the rising cost and complexity of developing large-scale AI models, even for companies with deep in-house research and silicon design teams.

How it works

The companies said Apple Foundation Models will be based on Google’s Gemini models and cloud technology. They did not describe the technical architecture or the operational division between the two firms. They also did not provide detail on which Gemini model variants Apple will use, or how often they will update.

The statement said Apple Intelligence will continue to run on Apple devices and on Apple’s Private Cloud Compute. That suggests Apple intends to keep its current approach that splits processing between local execution and remote compute under Apple’s control. The companies did not say whether Gemini-based components will run on Apple-controlled infrastructure, Google infrastructure, or a mix.

Apple and Google also said the arrangement will maintain Apple’s “industry-leading privacy standards”. The statement did not offer detail on data handling, retention, or the operational controls that govern the use of cloud resources.

Siri focus

The companies pointed to Siri as a near-term focus. The statement said future Apple Intelligence features will include “a more personalized Siri coming this year”. Apple has positioned Siri as a central interface for on-device and system-level interactions, with recent emphasis across the industry on assistants that can handle multi-step requests across apps.

The statement did not describe what “more personalized” will mean in practice. It also did not specify which languages, regions, or devices will receive the update first.

Evaluation process

The joint statement said Apple selected Google after an assessment of options. It indicated that Apple weighed different AI technologies before choosing Gemini as the basis for its next generation of foundation models.

“After careful evaluation, Apple determined that Google’s Al technology provides the most capable foundation for Apple Foundation Models and is excited about the innovative new experiences it will unlock for Apple users,” said Apple and Google in a joint statement.

Implications

The agreement may have implications for how Apple balances in-house model development with partnerships. Apple has invested heavily in custom silicon and machine learning frameworks across its operating systems. The statement suggests the company will anchor at least one layer of its next foundation models on Google’s technology.

For Google, the deal places Gemini at the core of a rival’s device ecosystem. It also puts Google’s cloud technology into the picture for future Apple Intelligence features, alongside Apple’s own infrastructure approach through Private Cloud Compute.

The companies did not disclose financial terms, commercial arrangements, or governance. They also did not say how the collaboration interacts with Apple’s other AI-related partnerships, developer tools, or third-party model access for applications.

The statement framed the collaboration as multi-year, which indicates a longer planning horizon. It also implies that model updates and product integration will span multiple software cycles and hardware generations.

Apple Intelligence will continue to run on Apple devices and Private Cloud Compute, while maintaining Apple’s industry-leading privacy standards.