Google's annual developer conference, I/O 2026, kicked off on May 19 with a focus squarely on artificial intelligence, signaling the company's entry into what it terms the "Agentic era." CEO Sundar Pichai emphasized the rapid advancements in AI, describing the period as "hyper progress" that is fundamentally transforming technology and its real-world applications.
During the keynote, Google showcased how AI is moving beyond experimental stages to become a core component of everyday user experiences. This shift is supported by substantial growth in AI adoption and significant infrastructure investments.
The Rise of Agentic AI and Usage Growth
Pichai revealed staggering growth in AI usage across Google's ecosystem. The company now processes over 3.2 quadrillion tokens per month, a substantial increase from 480 trillion tokens reported at Google I/O 2025. This exponential growth is driven by the integration of AI features into more products, with AI Overviews reaching 2.5 billion monthly active users and AI Mode surpassing 1 billion. The Gemini app alone has seen its active user base surge from 400 million to 900 million in just one year.
To support this escalating demand, Google has committed between $180 to $190 billion in infrastructure investment, with a significant portion allocated to its custom Tensor Processing Units (TPUs). At I/O 2026, Google introduced two eighth-generation TPU architectures: the TPU 8t, designed for large-scale AI training, and the TPU 8i, optimized for inference workloads.
Next-Generation Gemini Models and Tools
Gemini Omni and Flash for Video
A major announcement was Gemini Omni, a new family of multimodal models specifically engineered for advanced AI video generation. This model supports conversational editing, allowing users to modify characters, backgrounds, and other video elements using voice commands. Google is initially rolling out Gemini Omni Flash, a version optimized for AI video creation, with plans to expand its capabilities to image and text generation in the future.
Gemini 3.5 Flash and Antigravity 2.0
Google also launched Gemini 3.5 Flash, its next-generation flagship AI model. This model promises improved performance across coding, reasoning, and multimodal understanding, boasting four times faster performance compared to previous iterations and competing AI models. Complementing this, Google unveiled Antigravity 2.0, an enhanced agentic coding application. This app enables users to run specialized AI agents, create custom "subagent" workflows, automate tasks, and schedule processes to operate in the background without constant user input.
Revamped Gemini App and Personal Agents
The Gemini app received a significant visual overhaul, featuring a new "Neural Expressive design" with smoother animations, vibrant colors, updated typography, and haptic feedback. This revamped interface integrates the enhanced performance of Gemini 3.5 Flash, the AI video generation capabilities of Gemini Omni, and new agentic features powered by Gemini Spark.
Speaking of Gemini Spark, Google introduced this new personal AI agent designed to run 24/7 on behalf of users. Gemini Spark can be integrated within Google Workspace tools, enabling it to gain context from user activities and execute tasks autonomously in the background.
Beyond Software: Google's Hardware AI Push
Google also offered a preview of new audio-powered smart glasses, slated for release later this year. Developed in partnership with Warby Parker and Gentle Monster, these smart glasses are designed to function similarly to Meta's Ray-Ban glasses but will be powered by Google's Gemini AI model, extending AI's reach into wearable technology.
Broader AI Integration
Further demonstrating its commitment to pervasive AI, Google showcased Gemini for Science, a suite of AI tools tailored for researchers. Additionally, new AI-powered features are coming to popular platforms like YouTube, including "Ask YouTube" for video search, and "Docs Live" for seamless documentation generation, streamlining content interaction and creation.
This year's I/O underscored Google's profound transformation towards an agentic AI ecosystem. The vision is clear: intelligent systems will increasingly assist users, automate workflows, and operate across various applications, devices, and services with minimal human intervention, making AI an indispensable part of daily digital life.