Google I/O 2025: All You Should Know

The 2025 edition of Google I/O commenced with a video introduction generated by AI, set to the tune of New Radicals’ “You Get What You Give,” and concluded nearly two hours later with CEO Sundar Pichai’s final thoughts. The event was filled with announcements, updates, and product launches—particularly in the realm of AI.

Among the key features: Google’s Gemini AI model continues to shine, leading the LMArena leaderboard and even completing a playthrough of Pokémon Blue. However, the real emphasis was on the future of Google’s products and platforms.

Below is a detailed summary of the most significant announcements from Google I/O 2025—so you can skip the full 117-minute keynote.

Surprising Pricing for Google AI Ultra

One of the biggest surprises came early on: the cost of Google’s new AI subscription levels. While the standard plan starts at $19.99/month, the premium Google AI Ultra plan carries a hefty $249.99/month price tag. This plan provides full access to the complete array of AI tools with the most generous usage limits.

Google Search Becomes Smarter with AI Mode

Arguably the most significant update is the introduction of AI Mode in Google Search, now accessible to all users in the U.S. This feature enables more intricate, multi-part queries using a “query fan-out” strategy. Google mentions that AI Mode even “verifies its answers,” although specifics on the process remain murky.

Arriving later this summer, AI Mode will enhance its capabilities with:

– Personal Context: Customized suggestions based on your prior interactions with Google products like Gmail.
– Deep Search: An in-depth exploration of subjects by examining multiple websites.
– Data Visualizations: Charts and visuals generated from search results when relevant.

With 1.5 billion users already perusing AI summaries in Search each month, AI Mode has the potential to become one of Google’s most popular features.

AI-Enhanced Shopping Tools

Google also announced significant enhancements to its shopping platform. By leveraging the Shopping Graph, now encompassing over 50 billion products, users can articulate their needs and obtain customized results.

A notable demonstration included a presenter uploading her photo to virtually try on a dress—an AI-enabled feature now available in Google Labs. Additionally, users can utilize an AI shopping agent to monitor product availability and pricing changes.

Android XR and Project Moohan

Google finally unveiled its vision for AR/VR with the launch of Android XR, a new operating system tailored for immersive devices. The company highlighted two main products:

– Project Moohan: A headset developed in collaboration with Samsung and Qualcomm, expected to launch later this year.
– Android XR Glasses: Smart glasses equipped with built-in cameras, speakers, and displays. These will be trendier compared to the original Google Glass due to partnerships with Gentle Monster and Warby Parker. Developer access will begin next year.

Gemini Takes the Spotlight

Google’s prominent AI model, Gemini, was the highlight of the event. Notable updates include:

– Gemini 2.5 Pro: The latest version, capable of transforming sketches into complete applications.
– Gemini 2.5 Flash: A more budget-friendly variant that will debut in early June.
– Gemini 2.5 Pro Deep Think: Tailored for complex mathematics and coding, initially available to select testers.

Other key announcements related to Gemini:

– Jules: An asynchronous coding assistant currently in public beta.
– Native Audio Output: A text-to-speech model that retains the same voice across various languages.
– Agent Mode: A new feature in the Gemini app that functions as a task-completing AI assistant.
– Personalized Smart Replies: Launching in Gmail this summer, this feature utilizes personal data to generate context-sensitive responses.
– Deep Research: Enables users to upload files to direct AI outputs.
– Gemini in Chrome: Provides contextual support based on the current webpage, starting this week for U.S. users.

Google aims to incorporate Gemini across all its devices, ranging from smartwatches to smart TVs.

Generative AI: Imagen 4, Veo 3, and Beyond

Google also revealed an array of new generative AI tools:

– Imagen 4: An advanced image generation model with enhanced detail and typography.
– Veo 3: A video generation model offering improved physics and sound production.
– Flow: A video creation tool powered by Veo 3.
– Lyria 2: A model for generating music.
– Gemini Canvas: A collaborative space for co-creating content using AI.

Project Starline Transformed into Google Beam

Google’s immersive communication initiative, formerly called Project Starline, has now been rebranded as Google Beam.

Key features encompass:

– Google Meet Translations: Real-time translation of speech that replicates the speaker’s voice and tone. Currently available in English and Spanish.
– 3D Conferencing: A prototype system utilizing multiple cameras to produce a 3D light-field display of participants.

Project Astra and Gemini Live