Google I/O 2025 is officially underway. From AI-powered search and wearables to a new universal assistant and next-gen Android XR devices, Google is doubling down on making artificial intelligence the core of everything.
Here’s a complete breakdown of the biggest announcements (so far) from the event.
Gemini 2.5 Arrives
Google unveiled Gemini 2.5, the most advanced version of its flagship AI model yet.
A new capability called “Deep Thinking,” which mimics human-like reasoning by analyzing complex problems step-by-step.
Key highlights:
- More accurate reasoning
- Seamless multimodal input (text, image, audio)
- Now integrated into Search, Workspace, and Android
This version of Gemini is now powering features across Google, including the next-generation AI in Google Search, Workspace tools, and mobile devices.
Project Astra: The Vision of a Universal AI Assistant
Project Astra represents Google’s ambition to create a universal AI assistant capable of real-time multimodal interactions.
Demonstrated through a live, unedited video, Astra showcased its ability to understand and respond to visual and auditory inputs, such as identifying objects, explaining code, and recognizing locations.
Google says it’s bringing Astra to Pixel devices later this year, starting with the Gemini Live voice assistant.
Android XR and Smart Glasses
After years of teasing, Google officially announced a new Android XR platform, built specifically for extended reality (XR) devices.
In partnership with Samsung and Qualcomm, Google is developing a new wave of smart glasses and XR wearables, including
- Android-powered smart glasses with Gemini AI integration
- A new XR headset from Samsung, codenamed Project Moohan, set for release in late 2025
Although details are still sparse, demos hint at real-time translation, object recognition, and AI-driven AR overlays.
Search Gets an AI Mode
Google is giving Search its biggest update in years with a new AI Overviews feature.
Now, when users ask complex questions, Search will automatically generate a comprehensive summary using Gemini’s reasoning power.
AI Overviews are now rolling out to U.S. users and coming to over 1 billion people by the end of 2025.
Veo and Imagen 4
Two new creative AI models are also making headlines:
- Veo: A video-generation model that can create high-quality cinematic clips from text prompts.
- Imagen 4: A new version of Google’s text-to-image tool that can generate ultra-realistic photos.
These tools are launching inside Google’s experimental AI Playground and will eventually integrate into YouTube and other creative platforms.
AI Premium Plans
Google launched a new AI Premium subscription tier that includes
- Access to Gemini Ultra 1.5
- Integration across Gmail, Docs, and more
- AI agent capabilities with memory and history
It’s aimed at developers, creatives, and business pros who want cutting-edge tools, priced at $250/month.
As Google I/O 2025 continues, more announcements are anticipated, further detailing Google’s roadmap for integrating AI into every facet of technology.
