Google’s I/O announcements signal a deeper entrenchment of Gemini across the smartphone experience. This blog examines the harder question is where Android vendors (outside Mainland China) can still hold ground.
The Structural Realities I/O Laid Bare for Android Vendors
Worth noting, with Google services absent in Mainland China, the AI landscape there operates under entirely different conditions. This analysis will focus on the global market outside Mainland China. Google’s I/O announcements point to three structural advantages that Android vendors cannot easily replicate.
First is the application layer. Google can add Gemini to any app it already owns and reach every Android user instantly. Ask Maps is the clearest example. Google Maps is the default mapping app for almost every Android user outside Mainland China, and no OEM has a mapping app with anything close to that adoption. When Google builds AI into Maps, it lands on hundreds of millions of phones overnight, and no OEM can answer it.
The second is the identity layer, Google knows the user across every device they own, because they are signed in to one Google account on their phone, laptop, browser and tablet. Gemini draws on everything tied to that account, no matter which device the user is on. An OEM’s account, by contrast, mostly reaches its own hardware. Search autocomplete getting smarter with a signed-in account looks minor, but it works because Google sees the user everywhere, and the OEM does not.
The third is an orchestration layer. Gemini Spark does not just sit inside one app. It works across many of them at once. At I/O, Spark was shown reading an email, checking the calendar for a clash, and booking a meeting, all in one chain, without the user opening any app. The value is in connecting those apps, not in any single one. Spark can do this because Google also builds the apps, the operating system and the account they all run on. An OEM trying to match it has to connect apps it does not own, one integration at a time, and that is the gap Google’s software background makes hard to close.
Why Google’s Position Is So Hard to Contest
Outside Mainland China, Google and Gemini are present at every layer of the smartphone AI stack, from foundational model to agent layer.
Google’s entrenchment runs deep across the software and services that the internet economy relies on. It is a cloud hyperscaler and a core pillar of the digital world, with tools like Google Analytics and Google Ads Manager doing the tracking and attribution that e-commerce and online advertising run on. Google’s adjacent businesses are what make heavy investment in cutting-edge AI R&D worthwhile, unlocking synergies not available to others. Universal Cart is a clear example, a natural fit alongside Google’s existing e-commerce and online advertising operations. For smartphone vendors, whose business scope is far narrower, that kind of investment makes little sense.
This dynamic makes the relationship between Google and Android OEMs more complementary than competitive. Google has the Pixel but realistically needs Android OEMs for distribution at scale globally. OEMs gain world-class AI infrastructure without building foundational models from scratch. The strategic tension sits underneath that mutual benefit, and it is most visible when looking at Samsung specifically.
Samsung As the Test Case for Android Vendors Globally (outside Mainland China)
The picture painted above is a difficult one for Android vendors. But difficult is not the same as hopeless. Samsung is the most instructive case study for how a vendor can still carve out strategic ground within Google’s ecosystem.
As the global market leader in Android smartphones, Samsung holds more leverage with Google than any other Android vendor. It also has the most to lose if it cedes the AI layer entirely, because unlike Chinese OEMs, it has little presence in the China market where Google is absent to fall back on. Xiaomi’s internet services segment generated $5.2 billion in 2025 at margins above 75%, the majority from Mainland China. Samsung’s internet services revenue sits below 3% of total corporate revenue. The software layer is exactly where Samsung is most exposed.
Where Samsung Can Still Hold Ground: On-device Data
Samsung cannot out-build Google in software. But it can hold strategic choke points that Google has to go through Samsung to reach. The most defensible is on-device data from its own hardware ecosystem. Galaxy Watch and Samsung Health sit on years of longitudinal data across sleep, stress, heart rate variability, and activity. That data does not live in Gmail or Google Drive. Gemini Spark cannot access it without Samsung’s cooperation. A planning or health agent built on top of that data, reasoning across wearable signals in ways Google’s app-bound agents cannot replicate, is a genuine product differentiator that compounds over time as the dataset deepens. Samsung does not need to win the software layer broadly. It needs to own the layers where its hardware generates data Google cannot see.
Samsung’s partnership with Perplexity on the Galaxy S26 points in the same direction. Perplexity is integrated at the OS level, powering Bixby’s search and reasoning alongside Gemini. Samsung has not moved away from Google. But giving a second AI partner meaningful system-level access signals it is keeping its options open rather than consolidating entirely around one stack.
Hardware Distribution as Currency: Samsung’s Strongest Card
Samsung’s scale is its strongest card with Google. In 2025, it accounted for over 31% of Android shipments outside Mainland China and over 78% of the above-$700 segment, where AI features matter most. That makes Samsung the single most important partner for distributing Google’s mobile AI at scale.
A natural near-term collaboration is bundling free trials of Google One AI Premium with Galaxy device purchases, a model Android vendors including Xiaomi have already executed with other Google products. It drives subscription conversion, gives Samsung a tangible AI story at point of sale, and structures the Google relationship commercially rather than leaving value on the table.
It also surfaces an important question: are users buying a Galaxy device to access Gemini’s AI stack, or accessing Gemini’s AI stack because they bought a Galaxy device? The answer has direct implications for the commercial terms with Google, including cost allocation and potential revenue-sharing arrangements.
Trust as a Differentiator
According to Omdia’s consumer survey across five countries, 25% of respondents said the AI features they had seen would make them more concerned about their data and privacy. As agentic AI spreads, that concern only sharpens, because these systems work by reaching deeper into personal data. Google’s I/O launches are impressive, but each one routes more of the user’s life through Google, much of it invisibly.
Samsung’s in-house Knox gives it an advantage here, a security stack it controls outright and can build directly into how its AI handles user data. Samsung has already pointed this at AI with KEEP, which keeps the personal data behind features like Now Brief encrypted and on-device. The opening for Samsung is to build on that. As the company that owns the hardware and the security layer, it can make data protection a real point of differentiation, at a moment when users are telling vendors plainly that they are worried.
What Is at Stake and What Is Next? Different Vendors, Different Strategies
Google’s $20 billion annual revenue sharing agreement with Apple to be Safari’s default search defined a decade of mobile economics. The question for the smartphone AI era is what that arrangement looks like when the default is not a search engine but a persistent agent managing your inbox, calendar, and daily decisions. That question has not been answered yet. What is clear is that the right response will look different for every vendor, shaped by where they operate, what software infrastructure they already own, and what commercial objectives they are optimizing for. The vendors who navigate this well will be those who are clear-eyed about where their leverage sits.
