Google I/O 2026: Google Pushes AI Into Its Next, More Agentic Era

Google is taking its AI strategy to the next level. At Google I/O 2026, CEO Sundar Pichai outlined a series of major developments across Gemini, Google Search, AI agents, infrastructure, developer tools and generative media.

The central theme of this year’s event was clear: Google believes AI is moving beyond simply answering questions and generating content. The company is increasingly focused on AI agents that can understand goals, work across different tools and take action on a user’s behalf.

Google I/O 2026 Shows AI Adoption Is Accelerating

Pichai highlighted the rapid growth of AI usage across Google’s products and platforms.

Google says its systems now process more than 3.2 quadrillion tokens every month, a sevenfold increase from the previous year. More than 8.5 million developers are also building applications with Google’s AI models each month, while Google Cloud customers are processing enormous volumes of AI workloads.

The growth is also visible in Google’s consumer products.

AI Overviews in Google Search now have more than 2.5 billion monthly active users, while Google’s AI Mode has surpassed 1 billion monthly active users.

The Gemini app has seen particularly strong growth. Google says the service has passed 900 million monthly active users, compared with 400 million at last year’s I/O.

Gemini 3.5 Flash targets speed and intelligence

One of Google’s biggest announcements was Gemini 3.5 Flash, a new AI model designed to combine strong reasoning and coding capabilities with high output speed.

Google says Gemini 3.5 Flash performs better across almost all of its benchmarks compared with Gemini 3.1 Pro and has made significant progress in coding and real-world tasks.

Speed is another major focus. Google claims Gemini 3.5 Flash can produce output up to four times faster than other frontier models.

The company says the model is also significantly cheaper than comparable frontier models, potentially making advanced AI more accessible to developers and businesses.

Gemini 3.5 Flash is available across Google’s products and APIs, while Gemini 3.5 Pro is expected to arrive later.

Google introduces Gemini Spark

Google is also pushing deeper into agentic AI with Gemini Spark.

Unlike a traditional chatbot that waits for a prompt, Spark is designed to work in the background and complete longer-running tasks under the user’s direction.

Google says Spark runs on dedicated virtual machines in Google Cloud and can operate continuously without requiring the user’s computer to remain open.

The agent will integrate with Google’s tools and is expected to support third-party services through the Model Context Protocol (MCP).

Google is beginning to roll out Gemini Spark to trusted testers, with a beta planned for Google AI Ultra subscribers in the United States.

Search is becoming more agentic

Google also announced new AI capabilities for Search.

The company is introducing information agents designed to find information in the background and help users take action.

Search is also getting more interactive. Google says users will eventually be able to create custom experiences, dashboards and trackers directly within Search for longer-running tasks.

This represents a significant change in how Google views search. Instead of simply returning a list of links, Search is increasingly becoming an environment where AI can help users complete complex tasks.

Ask YouTube and voice-powered Docs

Google is also bringing more conversational AI into its everyday products.

Ask YouTube is designed to help users find relevant information within videos and jump directly to the sections that answer their questions. Google says the feature is currently being tested and will roll out more broadly in the United States this summer.

Another new feature, Docs Live, allows users to create documents through natural voice conversations. Instead of carefully writing a prompt, users can essentially talk through their ideas while Gemini organizes them into a document.

Google says voice capabilities will also expand to Gmail and Keep.

New AI infrastructure and Gemini Omni

Behind these products is a huge investment in computing infrastructure.

Google expects its annual capital expenditure to reach approximately $180–$190 billion in 2026, compared with $31 billion in 2022.

The company also highlighted its eighth-generation Tensor Processing Units, or TPUs. The new TPU 8t is designed for large-scale AI training, while TPU 8i focuses on inference and speed.

Google says TPU 8t provides nearly three times the raw computing power of its previous generation and can support training across more than one million TPUs globally.

Google also introduced Gemini Omni, a new model family designed to generate different types of outputs from different types of inputs. The first model, Gemini Omni Flash, focuses on video generation, with image and text capabilities planned over time.

Google expands AI transparency

As generative AI becomes increasingly realistic, Google is also putting more emphasis on identifying AI-generated content.

The company says its SynthID watermarking technology has already been used on more than 100 billion images and videos, as well as thousands of years of audio.

Google is expanding SynthID and Content Credentials verification to products including Search and Chrome.

The company also announced that OpenAI, Kakao and ElevenLabs are adopting SynthID, highlighting a growing effort across the industry to make AI-generated media easier to identify.

What Google I/O 2026 Means for the Future of AI

Google I/O 2026 shows that the company’s AI strategy is moving beyond standalone chatbots.

Gemini is increasingly being integrated into Search, YouTube, Docs, Android and other Google products, while AI agents such as Gemini Spark are designed to perform tasks rather than simply respond to questions.

For consumers, the biggest change may be the transition from asking AI for information to asking AI to get things done.

For developers and businesses, Google’s investments in models, agents and custom AI infrastructure could make increasingly capable AI systems faster and more affordable to deploy.

The next phase of the AI race may therefore be less about who has the best chatbot and more about which companies can build AI systems that are genuinely useful in everyday life.

Source: Google — Sundar Pichai’s Google I/O 2026 remarks
https://blog.google/innovation-and-ai/sundar-pichai-io-2026/