Google restricts Gemini to newer phones

Google said on May 12 that Gemini Intelligence will begin rolling out this summer on select Samsung Galaxy and Google Pixel phones, extending Gemini from a standalone assistant into what the company called an Android “intelligence system.” In blog posts for users and developers, Google described the new layer as handling multi-step tasks, summarizing web pages, filling forms and generating widgets, while stressing that the first wave is limited to its “most advanced” Android devices. Mindy Brooks, a Google vice president of product management, said Gemini Intelligence is designed to work “proactively” across the day while keeping users in control. Matthew McCullough, Google’s vice president of product management for Android Developer, told developers the shift creates new ways for apps to surface actions through the operating system rather than only through their own interfaces. (blog.google) ### Which phones did Google say will get Gemini Intelligence first? Google said the features will “roll out in waves” starting with the latest Samsung Galaxy and Google Pixel phones this summer. The company did not publish a broad consumer list in its main announcement, but it repeatedly narrowed availability to select devices rather than Android phones generally. (blog.google) Sameer Samat, president of Android Ecosystem, said on Google’s Android Show page that more Android updates will be discussed at I/O next week. Google separately said Gemini Intelligence will expand later this year to other Android form factors including watches, cars, glasses and laptops. ### Why are older premium phones at risk of missing the new features? (blog.google) Google’s official materials point to hardware as the dividing line. The company said Gemini Intelligence brings Gemini to its “most advanced devices,” and its developer documentation says on-device AI runs through AICore, the Android system service that manages Gemini Nano execution locally on supported hardware. (blog.google) Android Developers documentation says Gemini Nano is intended for on-device inference, with performance depending on device hardware. That matters because Google is pitching Gemini Intelligence as an OS-level feature set that can automate tasks and act on context across apps, not just answer prompts in a chat window. 9to5Google, citing a footnote on Google’s Gemini Intelligence page, reported that the requirements include a flagship chip, 12GB or more of RAM, AICore support and Gemini Nano v3 or higher. (blog.google) Reuters could not independently confirm the full footnote text from the pages reviewed, but Google’s own announcements and developer documents support the narrower point that availability depends on newer high-end hardware and AICore-backed on-device AI. (developer.android.com) ### How is this different from the regular Gemini app? Google’s Gemini mobile app remains broadly available on Android phones and tablets with much lighter requirements. Google’s help page says the Gemini mobile app works on Android tablets with 2 GB of RAM or more running Android 9 and up, and on Android phones including many foldables. That creates a split between basic Gemini access and the new Gemini Intelligence layer. (9to5google.com) Google’s support pages show Gemini as a widely distributed app and assistant, while the May 12 product announcements reserve the new proactive Android features for a smaller pool of devices. ### What are developers being asked to build around? Google told developers on May 12 that they can tap into the new system through AppFunctions, an Android API that lets apps expose services, data and actions to the operating system and AI agents. (support.google.com) McCullough said Google is also offering a “no-code change” path for some app automation, while testing early APIs in private preview with apps including KakaoTalk. The same post said AppFunctions have already enabled local execution of use cases across 25 apps and multiple device manufacturers. Google framed that as a way to let Gemini trigger app functions across phones, foldables, watches, cars and XR glasses. ### What comes next from Google? Google said on May 12 that Gemini Intelligence will start reaching select Samsung and Google phones this summer, with broader device categories following later in 2026. (developer.android.com) Samat said Google I/O, which begins May 19, will include more Android news and a preview of glasses that the company said will launch later this year. (blog.google)