Google launches Agent Compute Engine giving each Gemini agent its own microVM

Google’s Agent Compute Engine runs each AI agent in an isolated micro‑VM on the Titanium chip; idle agents pause, shared memory cuts cost, and tests with 200,000 agents showed suspend/resume under 500 ms. The service powers Gemini and will be available in preview in the coming months, supporting token tracking, model routing and third‑party identity providers.

Benchmarks are the company's own; not independently verified.

Why it matters

Agent Compute Engine’s 500 ms suspend/resume and shared‑memory cost cuts let product managers model pricing around per‑agent usage.

For product managers

For your product

What does this mean for your product?

Try it on your product

Read at The Decoder

More from this issueAll AI news from 23 Oct →

OpenAI seeks $30B raise at $1.25T valuation, sources sayUnconfirmed

OpenAI Makes Decisions API Widely Available, Challenging TypeSafe’s JevAvailable now

Anthropic releases Claude Duet proactive agent in research previewAnnounced

Microsoft's MAI-DxO AI system scores 83% accuracy vs 36% for doctors

Get this every morning, written for your product.

Deltaday reads the whole AI news cycle and tells you what changed, whether it's real yet, and what to do about it.

Request an invite See how it works