Google I/O 2025—Google’s flagship developer event—kicks off Tuesday and Wednesday at Mountain View’s Shoreline Amphitheatre. We’re reporting live with the latest from the show floor.
Across Android, Chrome, Search, YouTube and its AI initiatives—most notably the Gemini chatbot—Google has unveiled a slew of features and tools.
The Android Show
At this dedicated Android event, Google revealed enhanced device-finding capabilities, expanded protections under its Advanced Protection program, new security measures to thwart scams and theft, and introduced Material 3 Expressive, the next evolution of its design language.
Below is everything Google announced:
Gemini Ultra
Available now in the U.S. for $249.99/month, Gemini Ultra grants “top-tier” access to Google’s AI suite. Subscribers get:
- Veo 3 video generator
- Flow, a new AI-driven video editor
- Gemini 2.5 Pro Deep Think (coming soon)
- Increased quotas in NotebookLM and Whisk
- Chatbot access in Chrome plus experimental “agentic” tools from Project Mariner
- YouTube Premium and 30 TB of cloud storage across Drive, Photos, and Gmail
Deep Think (Gemini 2.5 Pro)
This advanced reasoning mode lets Gemini 2.5 Pro weigh multiple possible answers before delivering a response, improving benchmark performance. Currently in “trusted tester” preview via the Gemini API, Google is running extra safety checks before a wider launch.
Veo 3 Video AI
Veo 3 can generate video clips complete with sound effects, ambient audio, and dialogue—upgrading on Veo 2’s capabilities. It’s live now in the Gemini chatbot for AI Ultra subscribers, accepting prompts via text or image.
Imagen 4 Image AI
Googlers say Imagen 4 is significantly faster and will soon be up to 10× quicker than its predecessor. It excels at fine details—from fabrics to fur—and handles both photorealistic and abstract styles at up to 2K resolution. Imagen 4 and Veo 3 will power Flow, Google’s filmmaking-focused AI video tool.
Gemini App Updates
- Gemini Live now offers camera and screen-sharing on iOS and Android (powered by Project Astra), enabling near-real-time verbal chat with the AI.
- Soon, it will integrate with Maps, Calendar, and Tasks.
- Deep Research gains support for user-uploaded PDFs and images.
- Stitch: Generates web and mobile UI designs (with HTML/CSS output) from text or images.
- Jules: Expanded beta for bug fixes, code comprehension, and GitHub pull-request workflows.
Project Mariner
An AI agent that can browse websites and perform tasks—like buying event tickets or groceries—through a conversational interface, now updated to juggle nearly a dozen tasks simultaneously and rolling out to more users.
Project Astra
This low-latency, multimodal AI framework (originating from DeepMind) will power new Search experiences, the Gemini app, and third-party products. Google is co-developing Astra glasses with Samsung and Warby Parker, though no launch date is set.
AI Mode in Search
Launching this week in the U.S., AI Mode lets you ask complex, multipart Search queries via an AI chat interface. It’ll handle data-rich sports and finance questions and even “try-on” options for apparel. Coming later this summer, Search Live will let you ask questions based on your camera’s live view. Gmail is the first to get personalized context.
Beam 3D Teleconferencing
Formerly “Starline,” Beam combines a six-camera rig and a custom light-field display to create a 3D hologram–like video feed—complete with real-time speech translation that preserves voice and expression—integrated into Google Meet.
More AI Enhancements
- Gemini in Chrome: An AI browsing assistant for real-time page insights.
- Gemma 3n: A lightweight multimodal model previewed this week for phones, tablets, and laptops.
- Workspace AI: New Gmail smart replies, inbox-cleanup tools, Docs enhancements, and video editing features in Vids.
- NotebookLM Video Overviews: Summarize content in video form.
- SynthID Detector: Verify AI-generated media via Google’s watermark tech.
- Lyria RealTime API: Powers an experimental AI music-creation app.
Wear OS 6
Adds a unified font for cleaner tile designs and dynamic theming for Pixel Watch faces. Google is releasing design guidelines and Figma files to help developers craft smoother, more customizable watch experiences.
Google Play Developer Tools
New subscription-management capabilities, topic-based app pages, audio previews, and a revamped checkout experience. “Topic browse” pages (U.S. only) connect users to apps tied to specific shows and movies. Developers also get dedicated testing/release pages and the ability to pause live rollouts on critical issues.
Android Studio AI
Integrates “Journeys”—an agentic AI feature aligned with Gemini 2.5 Pro—and an “Agent Mode” for handling complex dev workflows. Enhanced crash insights, powered by Gemini, now analyze code to pinpoint and suggest fixes for stability issues.
That’s the full rundown of Google I/O 2025’s announcements. Stay tuned for deeper dives on all these innovations!