# Agentic infrastructure

Question: What infrastructure do AI agents need to run long, stateful work against private systems, safely and cheaply?

Status: ongoing. Started 2026-03-21. Last entry 2026-03-21. 1 entries, 12 min read, 4.0k tokens.

Agents that debug real systems need a place to run: a shell, tools, files, and network access to private APIs, for hours at a time. This journey is about that place.

## What we think now

Revision 1, 2026-03-21.

1. Sandboxes must run inside our own network. Hosted sandboxes offer that only on enterprise plans, from about $3K a month.
2. Persistence comes first. A sandbox has to survive timeouts and node loss; one 11-hour session survived a spot instance termination.
3. Design for idle time. Per 20 GB, a running sandbox costs about $5 a month, a paused one about $1.60 (resumes in about 5 seconds), an archived one about $0.50 (about 30 seconds). At 200 sandboxes with 80% idle, that is about $200 a month instead of $1,000.
4. Build on Kubernetes (OpenSandbox on EKS), with a network policy per sandbox and agents running unprivileged.

## Open questions

- Is container isolation enough, or do we need gVisor, Kata or Firecracker?
- Can the same sandbox run on a laptop?
- Can archiving and pre-warming be driven by usage numbers instead of fixed timers?

## Entries

1. [We Tried Every AI Sandbox. Then We Built Our Own.](https://engineering.cloudthinker.io/journeys/agentic-infrastructure/01.md) (2026-03-21, start, 12 min, 4.0k tokens)

Web version: https://engineering.cloudthinker.io/journeys/agentic-infrastructure/