Google I/O 2025: 10 Biggest AI Announcements in One Event

1. Gemini 2.0 Full Rollout

Google announced general availability of Gemini 2.0 across all tiers, with Gemini 2.0 Flash becoming the new default model in AI Studio and Google AI products. The 2.0 architecture, with native multimodal output and sub-200ms latency on the Live API, was the backbone of nearly every other announcement at I/O.

2. Project Astra Goes Live

Project Astra — Google's real-time multimodal AI assistant that can see through your phone camera and respond to the world around you — launched in limited beta for Pixel users. The assistant could identify objects, read text in the environment, answer questions about what it saw, and maintain context across a continuous real-world session.

3. NotebookLM Plus

Google launched NotebookLM Plus, a paid tier of its research assistant with higher source limits (50 sources, 500K words per notebook), team sharing, customisable Audio Overviews, and Google Workspace integration that could automatically ingest Google Drive files.

4. AI Mode in Google Search

AI Mode — a dedicated tab in Google Search that provided AI-synthesised answers to complex queries with cited sources — began a broader US rollout with global markets to follow. This represented Google's most direct response to Perplexity's AI search model.

5-10. The Supporting Cast

Additional announcements included: Veo 2 (video generation via Vertex AI), Imagen 3 (improved text-to-image), Google Lens real-time translation, Gemini in Gmail for full inbox management, Android AI features including Circle to Search improvements, and Google's new TPU v5p announcement for AI compute. The breadth of the announcements positioned Google as the only company with an end-to-end AI ecosystem spanning consumer, developer, and enterprise products.

What This Means for Indian Businesses