Edge AI Computing Box NVIDIA Jetson Orin NX Orin Nano Passive Cooling Computer

The 607 is an embedded edge AI computing box developed around the NVIDIA Jetson Orin NX and Orin Nano core boards. It brings up to 20, 40, 70 or 100 TOPS of AI compute to terminal devices and uses a fast passive cooling design with no fan, meeting industrial standards for shock resistance and anti-static protection. A wide set of interfaces and a cost-effective configuration make it suited to machine vision, robotics, smart cameras and edge inference deployments.

The 607 can be configured with four NVIDIA modules: Jetson Orin NX 8GB and 16GB delivering 70 and 100 TOPS with a 1024-core NVIDIA Ampere GPU and 32 Tensor Cores, and Jetson Orin Nano 4GB and 8GB delivering 20 and 40 TOPS with a 512-core or 1024-core Ampere GPU. CPU options are a 6-core or 8-core NVIDIA Arm Cortex-A78AE v8.2 64-bit CPU, and memory ranges from 4GB 64-bit to 16GB 128-bit LPDDR5. Storage is provided through an external NVMe drive on the M.2 Key M slot.

The interface panel provides two Gigabit Ethernet ports, four USB 3.0 Type A ports, one HDMI output, one CAN port with transceiver, one RS485 and one RS232 port, an SPI bus and an M.2 Key M socket. The unit accepts 12V (7A) to 36V (3A) DC input with a 70W power draw, operates from -25 to 60 C, and is housed in a 224 x 130 x 65.5 mm chassis weighing 2 kg.

• NVIDIA Jetson Orin platform: Built on Jetson Orin NX and Orin Nano core boards for up to 100 TOPS of edge AI performance.
• 20/40/70/100 TOPS options: Four module choices scale AI compute from 20 TOPS on Orin Nano 4GB up to 100 TOPS on Orin NX 16GB.
• Passive cooling: A fast passive cooling design removes fans for silent, dust-resistant 24/7 operation.
• Industrial durability: Designed to meet industrial standards for shock resistance and anti-static protection.
• Dual Gigabit LAN: Two 10/100/1000 Gigabit Ethernet ports support camera and sensor networks.
• Rich industrial I/O: CAN with transceiver, RS485, RS232, an SPI bus and four USB 3.0 ports connect industrial devices.
• HDMI display and M.2 storage: An HDMI output plus an M.2 Key M slot for external NVMe storage cover display and data needs.
• Wide-voltage input: A 12V (7A) to 36V (3A) DC input with 70W draw suits a range of power sources.
• Compact robust body: A 224 x 130 x 65.5 mm 2 kg chassis fits machine and cabinet installations.

The 607 edge AI box suits machine vision, robotics, autonomous systems, smart cameras and edge inference. Its NVIDIA Jetson Orin modules deliver 20 to 100 TOPS for real-time AI workloads, while dual Gigabit Ethernet, CAN, RS485, RS232 and four USB 3.0 ports connect cameras, sensors and industrial controllers. Passive cooling, a wide 12-36V input and a compact 2 kg chassis make it suitable for harsh industrial and embedded deployments.

Each unit ships in protective packaging suited to its 2 kg chassis. The NVIDIA Jetson Orin module, memory, external NVMe storage and interface options can be selected to match the target AI workload.