Polywell Computers: Nano-Jetson Series Mini PC

The main articles are:

• Computer
• Computers (Russian market)
• Computers (Global Market)
• Computers (CIS market)
• Computers (EMEA Market)

2025: Nano-Jetson Family Launch

On September 30, 2025, Polywell Computers announced the launch of the Nano-Jetson family, a class of ultra-compact systems built on the NVIDIA Jetson Orin platform. This marks Polywell's entry into the ARM + Jetson ecosystem, providing capabilities for AI inference, computer vision and robotics in a robust embedded form factor.

𝓲

Фото: Polywell Computers

According to the company, with the Nano-Jetson series, Polywell is expanding its horizons beyond x86/Intel-based industrial PCs, entering the fast-growing world of Jetson-based AI systems for edge computing. Now users building drones, robots, autonomous vehicles, smart factories and AI imaging devices can take advantage of Polywell, long-term support and industrial design - based on Jetson.

The main characteristics of the product Nano-Jetson 34/67/117/157:

• Processor/Chipset: ARM Cortex-A78 cores ( 6 or 8 cores) + Rockchip SoC RK3588
• AI/GPU/Inference: NVIDIA Jetson Orin Nano/NX modules with 34/67/117/157 TOPS performance (INT8)
• Memory: 4GB/8GB/16GB LPDDR5 (built-in, non-updatable)
• Storage/Extension: 1 × M.2-2280 (NVMe SSD slot) + TF card slot; additional AI accelerators or 5G modules via M.2 slots
• I/O and connection: Two gigabit LANs, built-in Wi-Fi/BT, 5 × USB, RS232/ RS485 ports, HDMI 2.1 (up to 8K 30Hz), GPIO, serial interfaces, MiniPCIe, SIM card slot, etc.
• Form Factor/Design: Fanless nano body (187 × 131 × 57 mm) with support for VESA or wall mount; individual colors are available
• Power and efficiency: 12V DC input (UL/CE/FCC certified), standby mode <1 Вт
• Operating conditions and reliability: Operation at -25 ° C to + 70 ° C, optimum shock and vibration resistance, MTBF 100,000 hours, full industry certificates (FCC, CE, RoHS)
• Support/ OS : ON 12, Debian Ubuntu 20.04 and up, with NVIDIA JetPack and all standard Jetson software stacks

Nano-Jetson capabilities:

• Jetson's native performance in a robust design: Unlike individual modules or reference media, Nano-Jetson offers a ready-to-deploy industrial-grade solution with reliable I/O interfaces, thermal design and support for advanced operating conditions.
• Flexibility on multiple levels: Choosing between 34, 67, 117 or 157 TOPS gives developers the flexibility to balance performance and power consumption on their system.
• Boundary AI inference: With built-in tensor cores and CUDA support, Nano-Jetson can handle object detection, segmentation, multi-camera vision, local output and LLM more - with minimal latency and no dependency. clouds
• A set of interfaces for robotics, automation and connection: With RS232/485, GPIO, serial interfaces, WiFi, expandable/ LTE, 5G Nano-Jetson is suitable for robotics, smart vehicles, peripheral gateways, machine vision and industrial automation systems.
• Polywell reliability + Jetson ecosystem: industrial heritage and support for Polywell, and full compatibility with Jetson's extensive developer ecosystem.

Potential use cases and markets:

• Drones/UAVs: Real-time vision, obstacle avoidance, navigation, payload processing.
• Autonomous robots/self-driving vehicles/autonomous mobile robots: Local AI processing for perception and decision-making.
• Smart vehicles/mobile platforms: Onboard output for environmental perception, mapping and awareness.
• Smart manufacturing/industrial automation: Defect checking, robotic vision, adaptive control.
• Edge gateways and AI nodes: Local analytics, merging data from sensors, security, pre-filtering data before downloading to the cloud.
• Intelligent video surveillance/camera nodes: Object detection, tracking, anomaly detection without constant streaming to the cloud.
• Kiosks/ AR, VR intelligent retail, interactive installations: AI and low-latency vision.

The recent NVIDIA Jetson Orin Nano "Super" update increases performance to 67 TOPS, up ~ 70% from previous versions (SiliconANGLE). In addition, the Jetson architecture offers full support for CUDA, TensorRT, and other deep learning frameworks.

By embedding Jetson Orin in its own secure system, Polywell allows users to skip the module integration stage and move directly to deployment - reducing development risks and time to market.

With the Nano-Jetson series, Polywell enters the niche of Jetson-based AI systems for boundary computing. This launch expands Polywell's reach in robotics, drones, smart vehicles and embedded AI sectors.