AVerMedia D133ONB Industrial Box PC — NVIDIA Jetson Orin NX/Orin Nano

The AVerMedia D133ONB is an ultra-compact industrial box PC built around NVIDIA Jetson Orin Nano and Orin NX modules, designed for engineers who need to get a computer-vision or edge AI application from prototype to field deployment fast, in a footprint that fits where a full-size industrial PC simply won’t. If your project needs dual-camera capture and real-time inferencing packed into a space smaller than a lunchbox, this is purpose-built for exactly that.

Key Features

• NVIDIA Jetson Orin Nano or Orin NX Module — the D133ONB ships with an Orin Nano 8GB module, while the closely related D133OXB variant offers Orin NX 8GB or 16GB for higher throughput; both are ready for computer-vision and edge AI inferencing out of the box, no separate carrier-board integration required.
• Genuinely Pocket-Sized — a 93 x 81.2 x 81.9mm chassis, built for deployments where every centimetre of cabinet or enclosure space is already spoken for.
• Dual MIPI Camera Inputs — 2x 4-lane MIPI CSI-2 connectors (22-pin FPC, 0.5mm pitch) for stereo vision, dual-angle coverage, or independent camera feeds without an external capture card.
• 256GB NVMe SSD Included — a pre-installed M.2 Key M 2280 NVMe drive ships standard, so there’s no separate storage purchase or installation step before you can start deploying.
• Display & Networking Built In — HDMI output at up to 4K30 (Orin Nano) or 4K60 (Orin NX), 1x Gigabit Ethernet (RJ-45), and an M.2 Key E slot for Wi-Fi, covering the connectivity a vision application needs without extra adapters.
• Industrial Expansion Header — a 20-pin header exposing 2x I2C, 1x UART, and 9x GPIO for sensor integration, actuator control, and custom hardware triggers alongside the vision pipeline.
• Wide Operating Range — rated from –25°C to 60°C, built for cabinets, kiosks, and outdoor enclosures that see real temperature swings, not just a climate-controlled office.
• RTC Battery & Health Monitoring — onboard real-time clock with MCU-monitored battery life, useful for unattended installations where you need to know the device is healthy before a technician has to drive out to check it.

From Prototype to Field-Deployed Vision System

AVerMedia positions the D133/DX33 box PC family explicitly for smart retail, smart camera, and smart city solutions — applications where a Jetson dev kit’s open board layout is a liability rather than a convenience. Exposed pins, an unshielded fan, and a bare PCB are fine on an engineer’s desk; they’re a support ticket waiting to happen once the device is bolted inside a kiosk or ceiling-mounted above a retail aisle. The D133ONB solves that by putting a fully validated Jetson module, its power delivery, storage, and I/O into a sealed enclosure that’s ready to mount and forget.

The dual MIPI CSI-2 camera inputs are the feature worth building a project around. Most compact edge AI boxes offer a single camera connector and force you to add a USB capture card for a second feed — sacrificing latency and adding a failure point. With two native 4-lane CSI-2 inputs, the D133ONB can run stereo depth estimation, cover two fields of view simultaneously, or run redundant capture for reliability, all processed on-device with no round-trip to an external server. Combined with the pre-installed 256GB NVMe SSD, you can start capturing footage, running inference, and logging results the moment the unit is powered on.

Being direct about what’s confirmed and what isn’t: AVerMedia does not publish a specific AI TOPS figure for the D133ONB on its own product pages — the number that matters is the Jetson module’s own published rating (NVIDIA’s Orin Nano and Orin NX modules span roughly 40–100 TOPS depending on configuration and Super Mode), and the module you select determines that ceiling. This is an inferencing and computer-vision deployment platform running the standard NVIDIA JetPack/L4T software stack — the same environment used across the wider Jetson ecosystem — so models trained and validated on a Jetson developer kit deploy to the D133ONB with minimal changes. For teams building machine-vision, AIoT, or robotics products who need reliable, repeatable, field-hardened hardware rather than a lab prototype wearing a plastic case, that combination of proven software compatibility and an industrial-grade enclosure is the whole value proposition.

Specifications

• Compute Module: NVIDIA Jetson Orin Nano 8GB (D133ONB) / Orin NX 8GB or 16GB (D133OXB variant)
• Storage: M.2 Key M 2280 NVMe, 256GB SSD pre-installed
• Display Output: 1x HDMI, up to 4K30 (Orin Nano) / 4K60 (Orin NX)
• Camera Inputs: 2x 4-lane MIPI CSI-2 (22-pin FPC, 0.5mm pitch)
• USB: 2x USB 3.2 Type-A, 1x USB 2.0 Type-C (recovery)
• Networking: 1x Gigabit Ethernet (RJ-45), M.2 Key E 2230 for Wi-Fi
• Expansion Header: 20-pin, 2x I2C, 1x UART, 9x GPIO
• Power Input: 12V/5A via 3.5mm screw terminal, 9–24V accepted range
• Power Consumption: Approx. 14.4–31.5W depending on module and load
• Operating Temperature: –25°C to 60°C
• Storage Temperature: –40°C to 85°C
• Dimensions: 93 x 81.2 x 81.9mm
• Weight: Approx. 500g
• RTC: Onboard real-time clock with MCU battery-life monitoring
• Operating System: NVIDIA JetPack / L4T (Ubuntu-based)
• Certifications: CE, FCC, KC
• Target Applications: Smart retail, smart camera systems, smart city, AIoT, machine vision, robotics