Google restricts Gemini Intelligence to flagships
- Google said on May 12 that Gemini Intelligence on Android will start on select Samsung Galaxy and Google Pixel phones this summer. - Google’s published device floor requires 12GB of RAM, Gemini Nano v3 or higher, AICore, and a “qualified flagship chip.” (9to5google.com) - Android’s Gemini Intelligence page says broader rollout across watches, cars, glasses and laptops is planned later in 2026. (blog.google)
Google on May 12 introduced Gemini Intelligence, a new Android feature set that automates tasks across apps, summarizes web pages in Chrome, rewrites spoken notes into cleaner text and builds widgets from natural-language prompts. The company said the rollout will begin this summer on the latest Samsung Galaxy and Google Pixel phones. Google described the package as “the best of Gemini” on its “most advanced devices,” signaling that availability will be limited at launch. (9to5google.com) ### Which Android phones are in the first wave? (blog.google) Google said the first rollout will reach “the latest Samsung Galaxy and Google Pixel phones” this summer. In the launch post, the company said it had spent months tuning multi-step automation on the Galaxy S26 and Pixel 10 using food-delivery and rideshare apps. Android’s public Gemini Intelligence page does not list a full compatibility table. Instead, Google frames the launch around its newest devices and says the broader experience will expand to other Android form factors later this year. (blog.google) ### What hardware does Google say a phone needs? Google’s published requirements, cited by outlets that reviewed the Android page’s footnotes, set a high bar: at least 12GB of RAM, support for AICore, Gemini Nano v3 or higher, and a “qualified flagship chip.” The same requirements also include long software support windows and device-quality targets, according to 9to5Google. (blog.google) Google’s developer documentation says Gemini Nano runs through Android’s AICore system service, which manages on-device inference and model updates while using device hardware for low-latency processing. (blog.google) Google also says inference speed depends on device hardware, even when prompts stay on the device. ### Why do those requirements exclude some expensive phones? The 12GB floor is only part of the filter. Several recent premium phones meet that memory number, but reports citing Google’s requirements say devices also need Gemini Nano v3, not earlier Nano versions. (9to5google.com) Android Headlines, citing the requirement details, reported that some devices including parts of the Pixel 9 line, Galaxy Z Fold 7 and Xiaomi 17 series use Gemini Nano v2, which would leave them outside the launch group despite premium pricing. (developer.android.com) That report attributed the device list to coverage by AssembleDebug. Google has not, in the material reviewed here, published a full official list of excluded phones. But the company’s own launch language points to a narrower first wave centered on its newest Pixel devices and Samsung’s latest Galaxy handsets. (9to5google.com) ### What is Google actually shipping with Gemini Intelligence? Google said Gemini Intelligence will handle multi-step actions across apps and the web, including booking rides, shopping, filling forms and pulling information from connected services. The company also highlighted Chrome summaries, a Gboard feature called Rambler that cleans up spoken drafts, and custom widget creation from plain-language requests. (androidheadlines.com) Developer documentation shows the on-device Gemini Nano stack already supports prompting, summarization, proofreading, rewriting, image description and speech recognition through ML Kit GenAI APIs. (blog.google) Google says those APIs run through AICore and are designed to keep user data local when features are processed on-device. ### What has Google said about limits and safety? Google’s Android and Gemini help pages continue to warn users to double-check Gemini responses and not rely on them for medical, legal, financial or other professional advice. (blog.google) That language remains in place even as Google expands Gemini from a chatbot into a more action-oriented Android layer. The developer documentation also says app makers remain responsible for the safety of their client apps and user experience when using GenAI APIs. (developer.android.com) That leaves device makers and software teams with both hardware thresholds and product-review obligations as Gemini Intelligence expands. ### What comes next in 2026? Google said Gemini Intelligence will roll out in waves starting this summer on select Samsung and Google phones. The company also said the experience will become available across watches, cars, glasses and laptops later in 2026, though it has not yet published device-by-device timing on the pages reviewed here. (support.google.com) (blog.google) (developer.android.com)