Gemini Spark runs on dedicated cloud VMs — the persistent personal agent moves from local extension to always-on cloud service

Google's Gemini Spark, the personal AI agent introduced at I/O, runs on dedicated virtual machines in Google Cloud and stays available 24/7 — even when the user's device is off. Spark is powered by Gemini 3.5 Flash via the full Antigravity pipeline, has cross-app access to the user's Gmail, Calendar, Drive, Photos, and YouTube history, and autonomously runs multi-step tasks on the user's behalf.

The architecture is the consolidation move. OpenAI's Operator runs as a browser session; Anthropic's Claude Cowork runs as enterprise workflow tooling. Spark's dedicated-VM persistence model is the strongest 'always-on assistant' commitment any major lab has made — the agent operates whether the user is logged in or not, with Google Cloud absorbing the infrastructure cost.

For the agent-surface-bifurcation argument we made yesterday, Spark hardens the consumer-personal-agent lane. The moat is the cross-app data graph plus 24/7 cloud presence; competitors that don't offer dedicated cloud VMs run a shallower product. Apple Intelligence, when it eventually ships a comparable surface, will likely follow Spark's architecture rather than Operator's.

Google — agentic Gemini era I/O 2026 → · Google Cloud — I/O 26 innovations → · BetaNews — Google I/O 2026 Gemini models →