Skip to content
NewA calmer, faster ZeroHour

Google Cloud launches Agent Compute Engine, one microVM per agent

Original title: Google Cloud Launches Agent Compute Engine: A MicroVM per AI Agent

MarkTechPostby Michal Sutter ()8 min read1
Part of a story covered by 2 sourcesGoogle previews Agent Compute Engine with one microVM per agentMerged summary & timeline
AI summary · x-ai/grok-4.7Summaries by Vani AI

Customer preview is months away; open-source Agent Substrate is already on GKE for non-production use.

Google Cloud announced Agent Compute Engine, which runs each AI agent session in a hardware-isolated microVM that holds no credentials. Google says it already powers Gemini as an agent, Antigravity, and internal evaluation and reinforcement-learning pipelines, and unveiled it at AGNTCon + MCPCon. Idle sandboxes are checkpointed to zero active resources. The Kubernetes counterpart, Agent Substrate, is open for non-production GKE use, with production GA allowlist-only.

Unconfirmed: Boot time, density, pricing and a firm preview date for Agent Compute Engine itself were not published.

Key facts
  • Guests receive only a session-scoped SPIFFE certificate and a short-lived identity token.
  • MicroVMs have no standard network interface, default gateway or direct VPC access.
  • Agent Substrate reported 500 activations per second under 500 ms across 200,000 agents.
  • Google also cites over 1,000 dormant agents per host and up to 10x density.
  • No ACE boot time, density, pricing or firm preview date was published.
Explain this AI news with Vani AIWhat was announced, what is new and why it matters, in a new chat

From the original article

8 min read at marktechpost.com

Google Cloud announced Agent Compute Engine, a compute service that gives every AI agent its own hardware-isolated microVM with no credentials inside. It already runs Gemini agents and enters preview in the coming months. The post Google Cloud Launches Agent Compute Engine: A MicroVM per AI Agent appeared first on MarkTechPost.

MarkTechPost

By Michal Sutter · marktechpost.com

Read the full article at MarkTechPost

A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.