Product Code: GTWY-039
Moxa UC-8200 Series (UC-8220-T-LX) Arm Cortex-A7 dual-core 1 GHz IIoT gate...
Product Code: MNHO-283
Advantech ARK-1251-S7A1, Intel® Core™ Ultra 5 125U, DDR5 5600 MHz Embedded...


The Multi-Site Edge AI Solution is purpose-built for enterprises operating distributed estates factories, warehouses, and remote sites that today run on separate systems, separate connectivity, and no unified visibility. Designed to bring AI-driven inspection and monitoring to every location while keeping a single central view, this solution deploys local edge AI at each site and aggregates the results into one cloud dashboard, making enterprise-scale AI practical without the cost and latency of centralizing raw data.
Anchored by an Advantech ARK-1251 edge AI computer built on Intel Core Ultra 7 with integrated GPU and NPU, the solution runs AI inference directly at each site defect detection, anomaly analysis, and quality inspection without sending raw video or sensor streams to the cloud. Local inference means decisions happen in milliseconds, bandwidth costs stay low, and each site keeps operating intelligently even if its connection to the cloud is temporarily lost.
Each site is built on a resilient local backbone: A Moxa EDS-510E managed switch connects the site's existing cameras, PLCs, and sensors into the edge AI computer, while a Moxa UC-8200 industrial IoT gateway handles protocol conversion, MQTT publishing, and secure data forwarding. A Teltonika RUTX50 5G router provides high-bandwidth WAN connectivity with built-in VPN, linking each site securely back to the central platform.
The central cloud layer aggregates every site into a single dashboard fleet-wide visibility, cross-site analytics, and unified alerting while closing the management loop in the other direction: AI models, configurations, and policies are deployed and updated centrally and pushed back down to every edge site. This two-way orchestration is what makes AI manageable at scale, turning dozens of independent sites into one coordinated, centrally governed intelligent estate.
Distributed edge AI ARK-1251 (Core Ultra 7, GPU + NPU) runs local inference at each site without cloud latency
Local AI inference defect detection, anomaly analysis, and quality inspection processed on-site in milliseconds
Central fleet visibility every site aggregated into a single cloud dashboard with cross-site analytics
Two-way model management AI models and configurations deployed and updated centrally, pushed to every edge site
Resilient site backbone Moxa EDS-510E managed switch connects cameras, PLCs, and sensors to the edge AI
Industrial IoT gateway Moxa UC-8200 handles protocol conversion, MQTT publishing, and secure forwarding
High-bandwidth 5G WAN Teltonika RUTX50 provides fast, reliable connectivity with built-in VPN per site
Bandwidth efficiency raw data stays at the edge; only results and telemetry are sent to the cloud
Offline resilience each site keeps operating intelligently even during cloud connectivity loss
Scalable architecture new sites integrate into the central platform without redesigning the core stack
| Component | Model / Detail |
|---|---|
| Edge AI Computer | Advantech ARK-1251 Intel Core Ultra 7, GPU + NPU |
| Cloud Gateway | Moxa UC-8200 Industrial IoT Gateway, MQTT, protocol conversion |
| 5G Router | Teltonika RUTX50 Industrial 5G, VPN, high-bandwidth WAN |
| Managed Switch | Moxa EDS-510E Gigabit managed, site backbone |
| AI Workloads | Defect detection, anomaly analysis, quality inspection |
| Inference | Local edge inference, low-latency, offline-capable |
| Cloud Protocol | MQTT / TLS results and telemetry northbound |
| Model Management | Central deployment and update of AI models to all edge sites |
| Connectivity | 5G WAN per site with built-in VPN |
| Security | VPN (IPsec / WireGuard), TLS, device authentication, role-based access |
Multi-site manufacturers running quality inspection across several plants, retail and logistics chains needing centralized monitoring of distributed warehouses, utilities and infrastructure operators managing remote sites at scale, food and beverage producers requiring consistent AI-based inspection across facilities, and any enterprise seeking to deploy and centrally govern edge AI across a distributed estate without centralizing raw data.