Google I/O 2025 what to expect: More of Gemini AI, less of Android 16

Google I/O

Google I/O 2025, set for May 20–21 at the Shoreline Amphitheatre in Mountain View, is shaping up to be a different kind of event —  not in the way most Android fans might expect. In a notable departure from tradition, Google has split off its usual Android announcements into a standalone event, The Android Show, slated for May 13.

The move reflects a broader shift in Google’s priorities, with artificial intelligence now firmly at the center of its strategy. This year’s I/O will place heavy emphasis on the Gemini AI ecosystem, along with next-gen projects like Astra and Veo 2. The main keynote kicks off on May 20 at 10 am PT and will be livestreamed on YouTube, the Google Developer site, and app.

Gemini AI is expected to headline the show. Google is teasing upgrades for Gemini Advanced subscribers that aim to boost personalization, productivity, and cross-app interaction. The enhancements likely include on-device task handling and deeper integration with system-level apps.

One key reveal might be Project Mariner, a Gemini-powered AI agent in early testing. Designed to run in browsers, it can reportedly interpret images, interact with web forms, and manipulate content in real time—hinting at a more proactive and visual future for Google’s assistant technologies.

Other likely highlights of the keynote include Project Astra, an ambitious real-time AI perception agent, and Google Veo 2, an advanced tool for AI-driven video generation and editing. Google is also expected to showcase smaller, faster LLMs optimized for on-device use—an answer to competitors like Apple and OpenAI who are pushing for privacy-centric AI.

By pushing Android into its own spotlight outside the I/O keynote, Google is sending a clear message: the age of Gemini is here. Expect a conference less about phones and more about the intelligent systems that will run them