Google Launches Gemini Spark: 24/7 Cloud AI Agent for Workspace Automation
The new agentic AI uses Google Cloud virtual machines to execute multi-step tasks autonomously, even when user devices are powered down.
Google has expanded access to Gemini Spark, an "always-on" personal AI agent designed to execute complex, multi-step tasks autonomously in the cloud. Launched at Google I/O 2026 on May 19, the tool marks a significant departure from reactive chatbots by operating independently of a user's active device.
Powered by the Gemini 3.5 Flash model, Spark utilizes virtual machines on Google Cloud to remain active regardless of the user's hardware status. The agent integrates deeply with Google Workspace—including Gmail, Docs, Sheets, and Slides—and supports external applications such as Canva, OpenTable, and Instacart via the Model Context Protocol (MCP). Access is currently limited to Google AI Ultra subscribers in the United States. Pricing for the service is $200 per month, though some sources indicate an entry tier available at $100 per month.
The Shift to Agentic AI
This launch represents Google's strategic pivot from generative AI—which primarily provides text-based responses—to "agentic AI." These systems are designed to take direct action within a digital workspace, handling workflows that previously required manual oversight. This move follows a broader industry race to develop autonomous agents capable of managing professional and personal logistics without constant human prompting.
Implications for the Digital Workspace
By functioning as a proactive "chief of staff," Gemini Spark significantly reduces the manual effort required to manage digital information. The primary advantage is that users no longer need to "babysit" the AI, as it can gather necessary information independently. However, the level of background access required for the agent to function 24/7 raises substantial privacy and security concerns regarding the continuous monitoring of a user's digital life.
Future Outlook
As Google rolls out this cloud-based autonomy, the industry will be watching for further integration of third-party apps and the potential expansion of the service to other global markets. While the core functionality of the Gemini 3.5 Flash-powered agent is now live for Ultra subscribers, the long-term impact on user privacy and the stability of autonomous cloud execution remain key points of scrutiny.