Google introduced the Gemini Spark agentic AI assistant during I/O 2026.

      Google has introduced Gemini Spark, a personal AI assistant that is always on and can accept tasks via a dedicated Gmail address, browse the web using Chrome, and function continuously without users needing to keep a laptop open. The announcement was made at the I/O 2026 developer conference, showcasing it as Google’s boldest endeavor in the competitive landscape of autonomous digital assistants.

      Gemini Spark is powered by Gemini 3.5 and utilizes the Antigravity agent harness, which is part of Google’s enhanced platform for developing agentic software. It operates on dedicated virtual machines hosted on Google Cloud, allowing it to perform lengthy tasks in the background without monopolizing a user’s device. CEO Sundar Pichai described it as a “personal AI agent that helps you navigate your digital life.”

      One of the standout features of this product is its deep integration with Google's ecosystem. Users can communicate with Spark directly through a specific Gmail address, similar to messaging a coworker, and the assistant can automatically gather context from Gmail, Google Docs, and other Workspace applications with no need for manual configuration. This seamless connectivity provides Google a structural edge over competitors whose agents depend on third-party integrations to access similar tools.

      Josh Woodward, VP of the Gemini app at Google Labs and AI Studio, has been overseeing the project. Under his leadership, the Gemini app now supports over 900 million users in 230 countries. Additionally, Spark accommodates the Model Context Protocol (MCP), an open standard that enables it to connect with various external services beyond Google’s offerings.

      On mobile devices, users will be able to track Spark’s activities via Android Halo, a new notification feature that displays real-time status updates at the top of the phone screen. Google indicates that Halo will be introduced later this year with Android 17, essentially turning the operating system into a hub for persistent AI agents.

      The launch comes amid a flurry of competing agentic products. OpenAI has recently combined ChatGPT and Codex into a unified agentic platform led by co-founder Greg Brockman. Meanwhile, Anthropic’s Claude Cowork allows users to delegate tasks to an AI that completes them by controlling a desktop computer, and Salesforce has enhanced Slackbot into an agentic system with over 30 new AI capabilities.

      Google believes its main advantage lies in the level of integration it offers. Since it controls the operating system, browser, email client, and cloud infrastructure, Spark can function across all these layers without the challenges that independent agents encounter when trying to connect different tools. Whether this advantage will be crucial remains to be seen as Spark transitions out of Google's internal testing phase.

      Gemini Spark is currently undergoing internal testing at Google. The company aims to launch it in beta for Google AI Ultra subscribers in the United States next week. The AI Ultra plan, which Google has announced will decrease from $250 to $100 per month, provides five times the usage limits of the existing AI Pro plan along with 20 terabytes of cloud storage and YouTube Premium.