{"slug": "arca-sophia-open-core-building-air-gapped-local-ai-inference-for-industrial-plc", "title": "Arca Sophia Open-Core: Building Air-Gapped Local AI Inference for Industrial SCADA/PLC Systems", "summary": "Arca Sophia Open-Core is an open-source reference architecture for running quantized LLMs locally on air-gapped industrial SCADA/PLC systems. The project addresses the trade-off between cloud latency and plant security by containerizing inference at the edge and adding a deterministic safety layer to prevent raw model output from directly controlling actuators.", "body_md": "#\nArca Sophia Open-Core: Building Air-Gapped Local AI Inference for Industrial SCADA/PLC Systems\n\nIntegrating Artificial Intelligence into Operational Technology (OT) and Industrial Control Systems (ICS) presents a fundamental engineering trade-off: **Cloud latency vs. Plant Security.**\n\nIn continuous industrial processing, sending telemetry to public cloud LLMs breaks air-gapped network designs and introduces non-deterministic network delays.\n\nTo solve this, we designed **Arca Sophia Open-Core**—an open-source (AGPLv3) reference architecture for executing quantized LLMs directly at the edge, fully containerized and bound by physical safety rules.\n\n##\nKey Architectural Challenges in Industrial Edge AI\n\n-\n**Strict Air-Gap Requirements**: OT networks cannot expose open WAN ports or send raw telemetry to third-party endpoints.\n-\n**Resource Constraints**: Edge nodes in industrial cabinets operate with capped hardware specifications (e.g., 4 vCPUs, 8GB RAM).\n-\n**Non-Deterministic Risk**: Raw LLM output cannot directly trigger PLC actuators without a deterministic safety layer.\n\n##\nThe Stack & System Architecture\n\nArca Sophia Open-Core resolves these constraints through a modular Docker setup:\n\n-\n**Inference Engine**: Runs 4-bit quantized GGUF models (e.g., `Qwen3-8B-Instruct`\n\n) via local backends without internet access.\n-\n**PLC Simulation Layer**: Isolated container simulating real-time SCADA telemetry and register reads.\n-\n**Deterministic Shield (SILIC-ETHIC)**: A rules-based containment layer that sanitizes model inferences against physical operational thresholds before any output is passed down the pipeline.", "url": "https://wpnews.pro/news/arca-sophia-open-core-building-air-gapped-local-ai-inference-for-industrial-plc", "canonical_source": "https://dev.to/matiasmeza_09/arca-sophia-open-core-building-air-gapped-local-ai-inference-for-industrial-scadaplc-systems-5917", "published_at": "2026-08-13 21:38:38+00:00", "updated_at": "2026-08-13 22:17:12.646372+00:00", "lang": "en", "topics": ["artificial-intelligence", "large-language-models", "ai-infrastructure", "ai-safety"], "entities": ["Arca Sophia Open-Core", "Qwen3-8B-Instruct", "SILIC-ETHIC"], "alternates": {"html": "https://wpnews.pro/news/arca-sophia-open-core-building-air-gapped-local-ai-inference-for-industrial-plc", "markdown": "https://wpnews.pro/news/arca-sophia-open-core-building-air-gapped-local-ai-inference-for-industrial-plc.md", "text": "https://wpnews.pro/news/arca-sophia-open-core-building-air-gapped-local-ai-inference-for-industrial-plc.txt", "jsonld": "https://wpnews.pro/news/arca-sophia-open-core-building-air-gapped-local-ai-inference-for-industrial-plc.jsonld"}}