Dev Tools · 1h ago
9 AI agents run on 2 CPU cores and 3.6GB RAM in a gym's back office
A fitness gym in China runs 9 specialized AI agents on a server with 2 CPU cores and 3.6GB RAM, using DeepSeek APIs for LLM inference. The agents handle tasks from fitness report interpretation to investor due diligence. The system uses identity files as both documentation and runtime configuration.
Meridian48 take
This is a practical demonstration that multi-agent AI systems can be deployed on minimal hardware with a solid orchestration layer and API calls, challenging the assumption that such systems require GPU clusters.
Read the full reporting
We Run 9 AI Agents on 2 CPU Cores and 3.6GB RAM: The Engineering Memoir →
DEV Community
multi-agent-systemsedge-ai