Description
ALPON X5 AI is a rugged, fanless edge AI computer powered by the Raspberry Pi Compute Module 5 and the DEEPX DX-M1 accelerator, delivering up to 25 TOPS of on-device AI inference.
It integrates NVMe SSD storage, dual Ethernet ports, 4G LTE/eSIM connectivity and seamless integration with ALPON Cloud for remote management and OTA updates. Built for industrial, robotics or smart IoT applications, this next-generation edge device enables real-time AI processing directly on-site. Winner of the CES® 2026 Best of Innovation award.
Performance at a glance
- 25 TOPSDEEPX DX-M1 NPU
- 16 GBLPDDR4X system memory
- −20 → +60 °CFanless, rugged
Highlights
- 25 TOPS NPU AI acceleratorDEEPX DX-M1, 4 GB on-chip memory.
- 16 GB system memoryLPDDR4X with NVMe SSD support.
- 2.4 GHz CPU clockQuad-core Cortex-A76.
- LTE / Wi-Fi / 2× GbENo external networking hardware required. Always-on industrial connectivity stack.
- −20 → 60 °COperating range, validated.
- 12-24 V DCIndustrial wide-range input.
- WatchdogReboots out of stuck states.
- Fleet managementCentralized monitoring and control at scale.
Technical specifications
| Compute | |
|---|---|
| AI accelerator | DEEPX DX-M1 · 25 TOPS · 4 GB on-chip memory |
| System memory | 16 GB LPDDR4X · NVMe SSD support |
| CPU | Quad-core Cortex-A76 · 2.4 GHz (Raspberry Pi Compute Module 5) |
| Connectivity | |
| Network | 4G LTE / eSIM · Wi-Fi · 2× Gigabit Ethernet |
| Cloud | ALPON Cloud - remote management, OTA updates and fleet monitoring |
| Industrial | |
| Operating temp | −20 °C to +60 °C |
| Power input | 12-24 V DC wide-range input |
| Reliability | Watchdog - reboots out of stuck states |
| Regulatory | |
| Certification | Certification in progress: CE, FCC, IC, UKCA, RoHS, WEEE, REACH |
Sixfab edge AI ecosystem
One NPU, one SDK, three form factors for real-world deployment.
-
AI HAT+ for Raspberry Pi 513 / 25 TOPS
AI acceleration for Raspberry Pi 5. Up to 25 TOPS of local AI performance.
-
Edge AI Expansion Board25 TOPS · LTE/5G
Add LTE/5G connectivity, NVMe local storage, and multi-camera support.
-
ALPON X5 AI25 TOPS · −20 to +60 °C
Fanless, rugged, always-online edge AI computer for fleets and distributed industrial sites.
How it compares
| Feature | AI development boards | Industrial AI PCs | ALPON X5 AI |
|---|---|---|---|
| AI inference / compute | Limited TOPS, varies by board | 25 TOPS NPU | 25 TOPS NPU |
| Built-in LTE connectivity | - | - | Built-in, no external modem |
| Cloud fleet management | - | - | Included, no third-party SaaS |
| Zero-touch provisioning & OTA | - | - | ✓ |
| Fanless rugged design | - | Partial | −20 °C to +60 °C |
| Low-power AI (vs GPU) | High power, active cooling needed | GPU-based (high power) | NPU, no active cooling |
| PoC to production | PoC only, not production-ready | - | Purpose-built |
| Open / extensible architecture | Limited | Limited | Raspberry Pi CM5 + GPIO/USB expansion |
Use cases
- ManufacturingDetect defects on a conveyor using a single industrial camera in real time. Train YOLOv8 with Ultralytics, compile for DEEPX NPU, and deploy via ALPON Cloud OTA updates.
- Smart citiesCount and classify vehicles, cyclists, and pedestrians from roadside edge units. Aggregate metadata over cellular networks while keeping raw video fully on-device.
- InfrastructureMonitor pumps, switchgear, and substations in unattended environments. Run anomaly detection locally and transmit only event-based alerts over LTE.
- MaintenanceAnalyze multimodal industrial signals to model normal operating behavior and detect statistical deviations in real time. Infer failure probability locally without streaming raw data.
What's in the box
- ALPON X5 AI device (25 TOPS · INT8)Fanless aluminum enclosure with the DEEPX DX-M1 NPU and Raspberry Pi CM5 integrated; nothing to assemble inside. Four external SMA connectors on the back panel are labeled L, L, G and W.×1
- External antennas: 2× LTE, 1× GNSS, 1× Wi-Fi (labels match the connectors)×4
- 27 W USB-C PD adapter with four interchangeable plug heads (US / EU / UK / AU)×1
Not included (sold separately)
wall mount bracket · DIN rail mount kit (35 mm per EN 60715) · IP67 outdoor antennas · DC terminal power source (12-24 V DC)
Frequently asked questions
What AI workloads does ALPON X5 AI support?
ALPON X5 AI is optimized for real-time Vision AI workloads, including object detection, classification, segmentation, anomaly detection, OCR, and pose estimation, running production-grade models such as YOLO and ResNet optimized for the DEEPX DX-M1 NPU. While it is not currently focused on LLM inference, this capability is on the DEEPX roadmap and will be supported as the underlying silicon ecosystem evolves, with Sixfab extending support accordingly.
Does it require cloud connectivity for AI inference?
No. ALPON X5 AI runs AI inference fully on-device. Cloud connectivity via ALPON Cloud is optional and used for remote management, OTA updates, and fleet monitoring, not for real-time inference.
Is it suitable for industrial environments?
Yes. ALPON X5 AI is designed for industrial deployment with support for wide temperature ranges, sealed enclosure options, and stable operation under continuous 24/7 workloads.
What is the difference between an NPU and a GPU?
A GPU is optimized for parallel processing and commonly used for AI training, but it consumes more power and requires active cooling. An NPU is designed specifically for inference, delivering higher efficiency, lower power consumption, and fanless operation - making it more suitable for industrial edge deployments.
