Free Premium Plan Offer: Activate a device prior to October 1st and get Premium free for one year! $2,995 value. Find out more »
Edge AI compute, purpose-built
Engineered for Workloads Beyond Rendering
Node Max runs the screen application and the AI that powers it on the same managed appliance — conversational kiosks, multi-camera vision, multilingual wayfinding, and heavy container workloads in one device.
- 50 TOPS Ryzen AI NPU
- AMD's XDNA 2 neural processing unit delivers up to 50 TOPS on the NPU and up to 80 TOPS across the full platform — enough to run multimodal LLMs, multi-stream computer vision, and concurrent inference workloads on the device itself.
- Quad 8K-Capable Display Output
- Four independent displays via HDMI 2.1, DisplayPort 2.1, and dual USB-C with DP 2.1 — each capable of 7680x4320 at 60 Hz. Drive video walls, multi-zone command centers, or four 4K60 video streams concurrently in hardware.
- Up to 64 GB DDR5 and 8 TB NVMe
- User-serviceable DDR5-5600 SO-DIMM dual-channel memory and two PCIe 4.0 x4 NVMe slots. Run large LLMs, host model weights locally, and cache content at the scale serious edge-AI workloads require — without compromise.
TelemetryOS Edge + AI Runtime Catalog
AI Workloads as Docker Sidecars, Managed by TelemetryOS
Node Max runs the same TelemetryOS Edge OS as Node Pro, with a validated catalog of language, vision, and speech runtimes that ship as Docker container sidecars to the TelemetryOS application — orchestrated by TelemetryOS Studio across your fleet.
- Container-native AI runtime catalog. Run llama.cpp, Ollama, ONNX Runtime, Whisper, and the rest of the open-source inference stack as Docker sidecar containers next to your TelemetryOS application. TelemetryOS validates the runtime configurations end-to-end so partners do not build the inference stack from scratch.
- On-device LLM inference. Mid-size open-weight language models run on-device with significant headroom for the vision pipeline, ASR, and the screen application. The model ecosystem moves fast — Node Max is sized for it.
- OCuLink PCIe 4.0 expansion. Four lanes of PCIe 4.0 to an external GPU enclosure. Start with the integrated Radeon 890M and 50 TOPS NPU, scale up to discrete GPU class when workloads demand it. No other Node device exposes external PCIe.
- Dual 2.5 GbE for network segmentation. One port on the camera/sensor VLAN, the other on management and content. Camera-driven computer vision pipelines stay isolated from TelemetryOS Studio traffic on the same device.
Regulated and Managed
- PHI- and gaming-compliant on-edge AI. Healthcare PHI and casino biometric data cannot transit cloud LLMs in most regulated jurisdictions. Node Max runs the LLM and the vision pipeline on the device, in a hardened TelemetryOS Edge environment, so the data never leaves the property.
- Fleet-wide managed appliance. Every Node Max is factory-provisioned, auto-registers with TelemetryOS Studio, and receives signed OTA updates for firmware, OS, runtime, applications, and AI container images — the same managed pipeline as Node Mini and Node Pro.
Enterprise edge I/O
A Compute Appliance That Drives Screens
Node Max collapses what was previously two boxes — an edge-AI inference appliance and a signage player — into a single managed device with the network, expansion, and peripheral I/O the workload demands.
- Dual 2.5 GbE Network Segmentation
- Two RJ-45 2.5 Gigabit Ethernet ports enable on-device network segmentation — separate camera VLAN traffic from TelemetryOS management and content delivery without external switches.
- OCuLink eGPU Expansion
- Four lanes of external PCIe 4.0 for discrete GPU enclosures, NVMe expansion, or capture cards. Only Node device with external PCIe.
- Rich USB I/O
- Two USB-C ports with DisplayPort 2.1 alt mode and PD 3.0, two USB 3.2 Gen 2 at 10 Gbps, three USB 2.0, and Micro SDXC — for cameras, touchscreens, payment peripherals, scanners, and industrial USB devices.
- Audio + Serial
- Two 3.5 mm jacks (line-out + earphone/mic) for voice interaction and PA integration. RS-232/RS-485 via WebSerial for commercial display control, industrial machines, and AV systems.
Full edge-AI platform
Everything Node Max Enables
On-device LLMs, multi-stream vision, container workloads, quad-display output, AMD Ryzen AI silicon, and the complete TelemetryOS managed platform — in one factory-order appliance.
- On-Device LLM Inference
- Run open-weight mid-size language models locally via llama.cpp, Ollama, or ONNX Runtime — conversational kiosks, RAG-backed product discovery, on-device translation, multilingual wayfinding.
- Multi-Stream Computer Vision
- Process 4-8 concurrent camera feeds for audience analytics, planogram compliance, queue measurement, safety monitoring, and behavior-driven content — all on-device with sub-50 ms latency.
- Container Workloads
- Docker sidecar containers for AI inference, MQTT brokers, content caches, protocol adapters — orchestrated alongside the TelemetryOS application on one managed device.
- Quad Display Output
- Four independent displays at up to 8K each, with hardware video decode across all outputs — video walls, multi-zone command centers, and large multi-panel installations.
- AMD Ryzen AI Silicon
- 12 cores / 24 threads of Zen 5, Radeon 890M GPU with RDNA 3.5, and the XDNA 2 NPU — the same silicon AMD targets for digital signage, machine vision, and edge AI in their Embedded P100 series.
- Full TelemetryOS Platform
- Every capability of Node Pro is on Node Max — TelemetryOS Edge OS, signed OTA, Studio fleet management, the full SDK, MQTT, WebSerial, remote diagnostics — plus the AI silicon underneath.
Technical specifications
Node Max Specifications
| Processor | AMD Ryzen AI 9 HX 370 — 4x Zen 5 + 8x Zen 5c (12 cores / 24 threads), up to 5.1 GHz boost |
| Graphics | AMD Radeon 890M, 16 RDNA 3.5 cores at 2.9 GHz, DisplayPort 2.1 |
| AI Engine | AMD XDNA 2 NPU — up to 50 TOPS NPU, up to 80 TOPS total platform |
| Memory | 8 / 16 / 32 / 64 GB DDR5-5600 SO-DIMM (dual channel, user-serviceable) |
| Storage | 2x M.2 2280 NVMe PCIe 4.0 x4 — up to 8 TB total (user-serviceable) |
| Display Output | Four independent displays, up to 7680x4320 (8K) at 60 Hz per output — 1x HDMI 2.1, 1x DisplayPort 2.1, 2x USB-C with DP 2.1 |
| Concurrent Video | Up to 4x 4K60 streams simultaneously (HW decode for H.264, H.265/HEVC, AV1, VP9) |
| Networking | 2x 2.5 Gigabit Ethernet (Realtek RTL8125BG), Wi-Fi 6/6E/7, Bluetooth |
| PCIe Expansion | 1x OCuLink PCIe 4.0 x4 (external GPU enclosures, PCIe peripherals) |
| USB | 2x USB 3.2 Gen 2 (10 Gbps), 2x USB-C with DP 2.1 / USB 3.2 Gen 2 / PD 3.0, 3x USB 2.0 |
| Serial | RS-232/RS-485 via USB-to-serial adapters (WebSerial) |
| Storage Expansion | Micro SDXC reader (up to 2 TB) |
| Audio | 2x 3.5mm audio jack (line-out + earphone/mic), HDA codec |
| Operating System | TelemetryOS Edge (Arch Linux) |
| Container Support | Docker sidecar containers; validated AI runtime catalog (llama.cpp, Ollama, ONNX Runtime, Whisper, and more) |
| IoT Protocols | MQTT broker connectivity |
| Power | DC 19 V / 6.32 A — 120 W external adapter; system TDP 65 W |
| Dimensions | 182 x 117.5 x 39.9 mm (~0.85 L) |
| Cooling | Active cooling — 95 mm long-life fan (30,000 hour rated) |
| Mounting | VESA 75 / 100 mm; Kensington security lock slot |
| Operating Temperature | 0 to 35 °C continuous |
| Availability | Factory order only — contact sales for configuration, lead times, and pricing |
Learn more about Edge AI on TelemetryOS
Specify Node Max for Your Deployment
Factory order only. Talk to our sales team about the right configuration, lead times, and edge-AI workloads Node Max is designed for.