Google open-sources AX, an agent runtime built to run billions of agent jobs on one cluster
JafarNajafov · x · 2026-09-23
Google open-sourced AX, an agent infrastructure designed to run billions of agent jobs on a single server cluster, per its README.
Key design points:
- Hard resource isolation: every agent runs in a sealed sandbox with strict limits on memory and compute, preventing runaway loops from piling up bills from paid model calls.
- Pre-loading: each agent's code folders and tools are loaded before startup, so nothing boots cold.
- Network lockdown: internet access is cut off except for explicitly approved domains.
- Pause/resume: agents can be frozen and restored later at the exact same execution step.
- Jobs are described declaratively in a single short text file.
The framework targets the operational reality that agents remember state, execute unreviewed code, and autonomously make repeated paid API calls — AX is Google's answer to running that safely at massive scale.
Related event: Google Open-Sources Agent Orchestration Runtime ax, Soaring on GitHub(3 posts)→
More from coding & agent
- Dev's 'software factory' burns 5-10B tokens a day with 32 parallel coding agents — BLUECOW009 · 2026-09-23
- Dev shares how he built app moderation with Jev and LLMs — amos_gyamfi · 2026-09-23
- Dev argues the real AI agent problem is unmeasurable requirements, not capability — algo_diver · 2026-09-23
- $/task beats token price: outputs are <5% of tokens in agentic coding — xeophon · 2026-09-23
- Rethinking multiplayer agent UX: shared artifacts may beat a single group-chat box — max__drake · 2026-09-23
- Dev: if your eval costs $20 for 100 runs, the problem set isn't hard enough — pvncher · 2026-09-23