SAN FRANCISCO – Google today unveiled Gemini Spark, a persistent AI agent that works around the clock—drafting emails, monitoring inboxes, assembling documents, and even making purchases—without requiring users to keep their devices awake. The announcement, made at Google I/O 2026, marks the company's boldest leap yet from conversational AI to autonomous task execution.
“We are in that part of the cycle where people want to see real value in the products they use on a day-to-day basis,” Sundar Pichai, CEO of Google and Alphabet, said during a pre-keynote briefing. With Spark, he argued, that value comes from an agent that never stops working. It operates in Google's cloud, so “you don't need to keep your laptop open to make sure it's running.”
What Gemini Spark Can Do
Unlike conventional AI assistants that activate only when prompted, Gemini Spark runs persistently on Google Cloud infrastructure. It uses the new Gemini 3.5 Flash model and what Google calls the “Antigravity agent harness”—the same system powering the company's internal developer tools. In practice, Spark can accept a complex instruction like “email my boss a status update pulling the latest figures from our shared spreadsheet and the project timeline in our Slides deck” and execute it across multiple Google applications without further input.
“When you use it, it almost feels like you're tossing things over your shoulder—Spark's catching them and gets the job done,” said Josh Woodward, VP of Google Labs, Gemini App, and AI Studio, during the briefing.
The agent can also monitor inboxes round the clock, surface urgent messages, and eventually make purchases using saved payment methods—all while the user's laptop is closed and phone is locked.
Background: The Race for Autonomous AI
Google's move comes at an inflection point for the tech industry. Microsoft, OpenAI, Anthropic, and Apple are all racing to build AI systems that not only converse but act—completing multi-step workflows with decreasing human supervision. The competition has intensified as users demand tangible productivity gains from AI tools. Spark is Google's answer: an agent that works in the background, never sleeps, and integrates deeply with its office suite.
“We are in that part of the cycle where people want to see real value in the products they use on a day-to-day basis,” Pichai reiterated, underscoring the urgency.
The product also raises urgent questions about trust, spending guardrails, and what happens when an AI agent misinterprets a user's intent. Google has not yet disclosed detailed safety protocols for purchase authorization.
What This Means for Users
Gemini Spark could redefine personal productivity by handling time-consuming administrative tasks autonomously. However, it also forces users to confront new levels of trust: ceding control over email drafting, inbox filtering, and financial transactions to an AI. The cloud-based architecture ensures the agent can work even when devices are off, but it also means data is processed remotely, raising privacy concerns.
Experts warn that despite safeguards, misinterpretation risks remain. “An AI that spends your money without explicit confirmation could lead to costly errors,” said Dr. Elena Marchetti, a digital ethics researcher at MIT. “Google must ensure robust guardrails are in place before wide release.”
Rollout Timeline
Spark will begin rolling out this week to a small group of trusted testers. A beta version is scheduled for Google AI Ultra subscribers in the United States next week. No word yet on international availability or pricing beyond the Ultra subscription tier.
For now, Google is positioning Spark as a premium service, likely to be bundled with its highest-tier AI plan. The company expects to gather feedback from testers before expanding access.
For more context on the competitive landscape, see the Background section. To understand the potential implications, read What This Means.