Google I/O spotlights Gemini direction

Google’s public framing for I/O 2026 is already centered on AI before the conference opens on May 19. Google said in its official save-the-date post that this year’s event will cover “latest AI breakthroughs and updates in products across the company, from Gemini to Android and more,” with the conference running May 19-20 at Shoreline Amphitheatre in Mountain View, California, and online. (blog.google) (io.google) The company’s published schedule adds detail to that message. Google has set its main keynote for May 19 from 10:00 a.m. to 11:45 a.m. Pacific, followed by a developer keynote at 1:30 p.m., while multiple technical sessions later the same day are explicitly organized around Google AI, Android 17 and agent-oriented development. (io.google) Google has also spent the week before I/O pushing Android updates that use Gemini as the organizing layer. In a May 12 Android post, Sameer Samat, president of the Android ecosystem, said Google was entering an “agentic Gemini era,” and said I/O would include “even more Android updates” plus “a sneak peek at glasses, which will launch later this year.” (blog.google) ### Why does Gemini look like the center of the event? Google’s own event materials put Gemini and AI across nearly every major track rather than isolating them in a single product announcement. The May 19 session list includes “What’s new in Google AI,” “Build next-gen AI experiences with Google AI Studio and Google Antigravity,” and Android sessions that describe agentic automation as part of the platform itself. (io.google) The Android team has used similar language in pre-I/O posts. Mindy Brooks, Google’s vice president of product management, wrote on May 12 that Android was “transitioning from an operating system into an intelligence system,” with Gemini Intelligence aimed at automating multi-step tasks, summarizing content and handling form filling. (blog.google) ### What is Google telling developers to build? I/O’s session descriptions show Google steering developers toward agents that work across tools, not just chat interfaces. A May 20 session titled “Agent-first workflows from prompt to production” says developers need a full lifecycle to deploy, scale and manage AI-native applications, while another session describes Google Antigravity as an “agent-first IDE.” (io.google) Chrome and Android sessions make the same point in more practical terms. One Chrome session says coding agents including Gemini CLI can be given direct access to a web app’s runtime through DevTools, and an Android 17 session says Google is using “agentic automation” to help users get more done faster. (blog.google) ### Where do Android 17 and cross-device software fit in? Android 17 is on the formal agenda, and Google’s wording suggests the release is being presented as part of a broader device network. The session “What’s new in Android” promises platform updates for media, camera, desktop and large-screen apps, while another session says Android 17 moves the ecosystem into an “Adaptive Everywhere” reality spanning phones, cars, living rooms and immersive environments. (io.google) Those descriptions line up with Google’s recent Android messaging. Brooks said Gemini Intelligence features will begin on the latest Samsung Galaxy and Google Pixel phones this summer and expand later this year to watches, cars, glasses and laptops. (io.google) ### Are smart glasses actually part of this week’s story? Google has not published a dedicated I/O keynote page spelling out a glasses launch, but two official posts point to glasses as part of the near-term Android XR roadmap. Samat said I/O would offer a “sneak peek” at glasses that “will launch later this year,” and Shahram Izadi, Google’s vice president and general manager for XR, wrote in December that partner AI glasses were due next year, meaning 2026. (io.google) Android XR remains part of the company’s wider platform pitch. Google introduced Android XR as an operating system for headsets and glasses in late 2024, and more recent Android XR posts have tied Gemini directly to those devices. (blog.google) ### Which names should readers watch on the schedule? Liz Reid, Google’s vice president and head of Search, is listed among the speakers for a May 20 session titled “Defining the agentic AI era.” Jeff Dean, Koray Kavukcuoglu and Josh Woodward are also listed for that session, putting senior Search, DeepMind and Gemini leadership in one slot. (blog.google) The next public markers are already fixed. Google’s main keynote begins on May 19 at 10:00 a.m. Pacific, the developer keynote starts at 1:30 p.m., and the Android 17 and Google AI platform sessions follow later that afternoon on the I/O schedule. (io.google 1) (io.google 2) (blog.google)