# MicroVM, shared VM or no VM: the hosting economics of personal AI agents

> Priced from published rate cards: dedicated microVMs, shared VMs, snapshot/resume and VM-less actors for Muse-, Instinct- and Fo-style personal agents.

Source: https://tokenomy.ai/blog/personal-agent-hosting-economics
Last updated: 2026-10-11
Publisher: Tokenomy — FinOps for AI
License: free to quote with attribution and a link to the source URL.

What it costs per user per month to keep a cloud-hosted personal agent alive, with every calculation shown and every source linked.

## Overview

What it costs per user per month to keep a cloud-hosted personal agent alive, with every calculation shown and every source linked.

- Instinct-sized sandbox: $95.90/month always-on vs $4.00 on-demand at E2B rates
- Muse's observed 2 vCPU / 7.7 GiB box: $162/month always-on at managed rates
- Self-hosted with snapshot/resume: ~$0.15 per user
- With cheap models, idle compute becomes 94% of the bill

## About Tokenomy

Tokenomy is the economic runtime for AI — the system of record for AI model pricing, plus free tools, independent research and runtime rails that meter, enforce and attribute every model call.

## Related pages

- [Free tools](https://tokenomy.ai/tools)
- [Research](https://tokenomy.ai/research)
- [Pricing Data API](https://tokenomy.ai/data-api)
- [Academy](https://tokenomy.ai/academy)
- [Blog](https://tokenomy.ai/blog)
- [Pricing](https://tokenomy.ai/pricing)
- [Why FinOps for AI](https://tokenomy.ai/finops-for-ai)