Google I/O 2025 is here, bringing together thousands of developers and tech enthusiasts to explore the latest trends from Mountain View. This edition confirms the changing of the guard: artificial intelligence and extended reality are the main focus of the event , shifting the spotlight away from Android, whose main new features were unveiled at its own event the previous week.
Artificial intelligence is taking the lead in a time of fierce competition, with the company poised to showcase its latest advancements in generative AI, multimodal assistants, and new integrations into everyday products. During his keynote address, Sundar Pichai, CEO of Google and Alphabet, outlined the roadmap that will guide consumer and development technology in the coming months, emphasizing an AI that not only responds but also acts and personalizes each user's experience.
Gemini 2.5 and Gemini Live: The leap towards a universal assistant
The cornerstone of this edition is undoubtedly Gemini 2.5 , Google's most advanced AI model, which reinforces its role as a leader in conversations, complex tasks, speech understanding, and visual interpretation. Alongside it, Gemini Live enables real-time interaction with the device's camera and screen, generating contextual responses and instant visual assistance, with translation capabilities and intelligent environmental analysis.
One of the most notable new features is the implementation of Project Astra in Gemini Live . This development takes AI a step further, as it not only interprets videos and images but also actively responds to what the user shows or shares on their screen. In this way, assistance goes far beyond simple conversation, facilitating visual queries and real-time translation of signs and text.
The Gemini Live assistant also combines native audio output and affective dialogue , detecting the user's mood in their voice and adapting its response accordingly. All of this is supported by an expanded context window, capable of handling large volumes of information to provide accurate analysis of both text and video.
Project Astra and Project Mariner: From Assistant to Digital Agent
In the applied research section, two initiatives stand out: Project Astra , a multimodal system that understands, sees, listens to, and assists the user with unprecedented proactivity, and Project Mariner , which marks a paradigm shift from assistant to digital agent. Thanks to Mariner, AI can navigate the web, interact with interfaces, complete forms, and perform tasks such as managing reservations or purchases , always under the user's control.
The new agent mode is gradually being integrated into Chrome, Search, and the Gemini app, allowing AI to take the lead in everyday tasks. A practical example is apartment searching: AI can scan listings, adjust filters, and schedule viewings. Additionally, with the Agent2Agent protocol, multiple agents can work together to solve complex problems.
These features aim to simplify the user's digital life by streamlining processes, automating routine tasks, and anticipating needs through a "teach and repeat" learning capability. These capabilities will soon be available to a wider audience, with priority access for developers and businesses.
Generative models: Veo 3, Image 4 and Lyria 2
Digital creativity has taken a leap forward with the introduction of Veo 3 and Imagen 4. Veo 3 allows for the creation of videos with integrated audio , achieving realistic lip-syncing and advanced control of movement or visual storytelling, while Imagen 4 takes image generation to new levels of detail and speed, offering formats up to 2K.
Both models can be used in the Gemini app, Flow, Vertex AI, and Workspace products, with enhanced tools like outpainting, camera control, and object manipulation in videos.
Lyria 2 joins the ranks of music AI models, available to composers and musicians on platforms like YouTube Shorts and Vertex AI. It allows users to create and edit melodies in real time, exploring sounds and styles under their creative control.
Android 16, Material 3 Expressive, and XR: Software and hardware evolve
In the Android realm, the focus shifts to user experience and AI integration. Material 3 Expressive reinterprets the system's design, delivering smoother transitions, depth effects, improved typography, and dynamic customization options through new color themes.
Android 16 introduces not only improvements in efficiency and connectivity, but also compatibility with AI features and extended reality. Android TV 16 also receives a new design and stronger support for Gemini, while Android XR advances as an operating system for glasses and immersive devices, with prototypes developed in collaboration with Samsung and Xreal.
Among the new developments in XR were experimental smart glasses and a mixed reality headset (Project Moohan) , promising applications ranging from instant translation to navigation and interactive entertainment. These devices are expected to reach the market throughout the year, ushering in a new era in the relationship between users and the digital environment.
Ultra Subscriptions, APIs, and Apps: The Ecosystem Expands
Google is reinforcing its commitment to subscriptions and advanced services with Ultra and Pro plans that offer priority access to the most powerful models, additional features, and early access trials. Key benefits include access to Veo 3, Imagen 4, and Gemini, as well as expanded storage, exclusive multimedia editing tools, and priority support for enterprise and creator products.
In addition, new Gemini Nano APIs and improvements to NotebookLM are being introduced , as well as the expansion of AI to products such as Google TV and Android Auto, enabling personalized and contextual experiences on any connected device.
Google I/O 2025 reinforces its transformation toward ubiquitous, proactive, and creative AI, which acts and integrates into every aspect of everyday life, from productivity and creativity to entertainment and learning. Android, while losing ground on the main stage, continues to evolve with AI integration and the exploration of new platforms such as extended reality. Innovation continues, and upcoming applications of artificial intelligence are already shaping the digital future.