After several monthsRumors and speculationsFinally, Apple and GoogleJoint statementApple has officially confirmed a multi-year partnership agreement with Google, under which Apple will use Google’s Gemini model and cloud technology as the core of its Foundation Models to drive a more personalized new version of Siri, expected to debut later this year, as well as other Apple Intelligence features.
In an effort to save Siri, Apple ultimately chose Gemini.
This statement confirms recent reports from CNBC and Bloomberg News, and Apple stated that after "careful evaluation," it determined that Google's AI technology provides the "most capable foundation" for its underlying models.
This means that the new version of Siri, expected later this year (possibly with iOS 26.4 or at WWDC 2026), will no longer rely solely on Apple's self-developed AI model for its "brain," but will instead incorporate the computing power and architecture of Google Gemini. This collaboration will address Siri's long-standing criticisms of not being smart enough and lacking comprehension, and unlock more innovative user experiences.
Privacy remains the bottom line: it will be safeguarded by the Private Cloud Compute mechanism.
Despite incorporating Google technology, Apple strongly emphasizes its commitment to privacy. The statement indicates that Apple Intelligence features will continue to run on Apple devices and its own Private Cloud Compute.
In other words, Apple uses Gemini as a basic "knowledge base" or "inference engine," but in terms of data processing and privacy protection, it still encapsulates and controls it through Apple's own architecture to ensure that users' personal data does not flow directly into Google servers without processing.
From competition to symbiosis: A $10 billion compromise?
This collaboration is seen by the market as a sign that Apple is acknowledging a "temporary lag" in the AI race. Back in March 2025, Apple announced a delay in the release of the new version of Siri, indicating that its self-developed model was not progressing as expected. According to previous market rumors, Apple may have to pay Google up to $1 billion annually in licensing fees because of this collaboration.
However, this collaboration may be a win-win situation for both Apple and Google. Google can consolidate its leading position in generative AI and gain access to the huge demand for inference from billions of iPhones, while Apple can quickly catch up in its AI experience and avoid falling behind in the competition with companies such as Samsung and OpenAI.
Analysis of viewpoints
As previously reported, Google used to pay Apple $200 billion annually to maintain its search engine as the default option in Safari. Now, the tables have turned, and Apple must pay Google to use its Gemini AI technology. (But the result was clearly that Google paid a larger amount.).
This demonstrates that in the era of generative AI, "computing power" and "models" have become the new infrastructure. While Apple boasts a vast ecosystem of iPhone and Mac products, and is powered by A-series/M-series chips, it clearly cannot catch up with Google and OpenAI in the short term when it comes to the training and iterative updates of large-scale models in the cloud. Therefore, choosing to ally with Google, while perhaps losing face, is the most pragmatic strategy to stem the bleeding.
This will undoubtedly be good news for consumers, as the previously delayed, more conversational, and capable Siri, which can handle complex tasks, will no longer be hampered by Apple's slow-moving in-house development. Whether this will make Apple more reliant on Google, or is merely a transitional measure before Apple's in-house model matures, will be a key focus to observe in the coming years.
