Google announced Gemini Spark at I/O 2026, an always-on agentic AI assistant built on Gemini 3.5 and the Antigravity harness. It runs on dedicated cloud VMs, integrates natively with Gmail and Workspace, and will be available to AI Ultra subscribers next week.[/tldr]
Google has unveiled Gemini Spark, an always-on personal AI agent that can receive tasks via a dedicated Gmail address, browse the web through Chrome, and work around the clock without requiring users to keep a laptop open. The company announced the product at its I/O 2026 developer conference on Monday, positioning it as its most ambitious move yet in the intensifying race to build autonomous digital assistants.
Spark is built on Gemini 3.5 and powered by the Antigravity agent harness, Google’s newly expanded platform for building agentic software. It runs on dedicated virtual machines on Google Cloud, which means it can execute long-running tasks in the background without tying up a user’s device. CEO Sundar Pichai described it as a “personal AI agent that helps you navigate your digital life.”
The product’s most distinctive feature is its tight integration with Google’s own ecosystem. Users can email Spark directly through a dedicated Gmail address, much as they would message a human colleague, and the agent can pull context from Gmail, Google Docs, and other Workspace applications without requiring manual setup. That out-of-the-box connectivity gives Google a structural advantage over rivals whose agents must rely on third-party integrations to access the same services.
Josh Woodward, VP of the Gemini app, Google Labs, and AI Studio, has been leading the product’s development. Under his stewardship, the Gemini app now serves more than 900 million users across 230 countries. Spark also supports the Model Context Protocol (MCP), an open standard that lets it connect to a wide range of external services beyond Google’s own suite.
The 💜 of EU tech
The latest rumblings from the EU tech scene, a story from our wise ol' founder Boris, and some questionable AI art. It's free, every week, in your inbox. Sign up now!
On mobile, users will be able to monitor what their agent is doing through Android Halo, a new notification layer that surfaces live status updates at the top of the phone screen. Google says Halo will arrive later this year with Android 17, effectively turning the operating system into a dashboard for persistent AI agents.
The launch follows a wave of competing agentic products. OpenAI recently merged ChatGPT and Codex into a single agentic platform under co-founder Greg Brockman, while Anthropic’s Claude Cowork lets users hand off tasks that the AI then completes by controlling a desktop computer. Salesforce has also joined the fray, transforming Slackbot into an agentic system with more than 30 new AI capabilities.
Where Google believes it can differentiate is depth of integration. Because it controls the operating system, the browser, the email client, and the cloud infrastructure, Spark can operate across all of those layers without the friction that standalone agents face when stitching together disparate tools. Whether that advantage proves decisive will depend on how well Spark performs once it leaves Google’s internal testing environment.
Gemini Spark is currently being tested internally at Google. The company plans to roll it out as a beta to Google AI Ultra subscribers in the United States next week. The AI Ultra tier, which Google also announced it is cutting from $250 to $100 per month, includes five times the usage limits of the existing AI Pro plan along with 20 terabytes of cloud storage and YouTube Premium.
Get the TNW newsletter
Get the most important tech news in your inbox each week.