Specifications
Platform
NVIDIA Jetson Thor (T5000 / T4000 Blackwell GPU module)
AI Performance
Jetson Thor Blackwell GPU with FP4/FP8 Transformer Engine
Software Stacks
NVIDIA Isaac (robotics), Metropolis (vision AI), JetPack 7.0
Memory
Up to 128GB 256-bit LPDDR5X
High-Speed Networking
1x 100GbE QSFP28, breakable to up to four 25GbE lanes
Additional LAN
1x 5GbE RJ45
Camera Inputs
8x GMSL2 camera interfaces
Display
1x HDMI 2.0
USB
4x USB 3.2 Gen1
Serial / Fieldbus
2x RS-232/422/485 with optional CANbus, digital I/O and audio
Expansion
3x M.2: M-Key NVMe, E-Key Wi-Fi/Bluetooth, B-Key 4G LTE/5G
SIM
Dual Nano-SIM support
Operating Temperature
-25C to +70C
Power Input
24V / 48V DC
Operating System
Ubuntu 24.04 LTS, Linux kernel 5.15
Variants
EAI-I351A and EAI-I351B
Pricing
Quote (vendor list price not published)
Stock
Available
Overview
The Lanner EAI-I351 is a rugged edge AI computer built on NVIDIA's Jetson Thor platform. Its defining feature among Thor-based systems is a 100GbE QSFP28 port that can be broken out into up to four 25GbE lanes, lifting the system well above the 2.5GbE and 10GbE connectivity typical of first-wave Jetson Thor boxes.
Alongside that uplink the platform pairs a 5GbE RJ45 port, eight GMSL2 camera inputs for low-latency multi-camera vision, HDMI 2.0, four USB 3.2 Gen1 ports and two configurable RS-232/422/485 interfaces with optional CANbus, digital I/O and audio. Three M.2 sockets cover NVMe storage, Wi-Fi/Bluetooth and 4G LTE/5G cellular with dual Nano-SIM, so the unit can be deployed as a self-contained field node.
Up to 128GB of 256-bit LPDDR5X memory addresses the large-model and multi-stream video workloads Jetson Thor was designed for. The system runs the NVIDIA AI software stack from JetPack 7.0 on Ubuntu 24.04 LTS, with support for Isaac for robotics and Metropolis for vision AI, across an industrial -25C to +70C operating range and 24V/48V DC input.
QS Compute supplies the Lanner EAI-I351A and EAI-I351B as quote-based B2B orders. Pricing is not published by the vendor; contact our team with your configuration and volume requirement.
Key Benefits
100GbE-class networking on an edge device — a QSFP28 port that breaks out to four 25GbE lanes removes the network ceiling that limits first-generation Jetson Thor systems. Eight GMSL2 camera inputs support multi-sensor vision pipelines without external acquisition hardware. 128GB LPDDR5X comfortably hosts large vision-language and robotics models locally. Industrial hardening — -25C to +70C operation and 24V/48V DC input suit factory, vehicle and outdoor cabinets. Triple M.2 expansion gives independent NVMe, wireless and cellular paths with dual Nano-SIM failover.
Applications
Autonomous mobile robots and AMRs, multi-camera machine vision inspection, smart manufacturing cells, autonomous systems and teleoperation, video analytics at the network edge, robotics research platforms requiring Isaac/ROS, and deployments where vision data volume exceeds 10GbE capacity.
Request a Quote — LANNER EAI-I351 — NVIDIA JETSON THOR EDGE AI COMPUTER WITH 100GBE QSFP28 AND 8X GMSL2
QS Compute — global B2B supply of AI computing hardware, edge AI systems and accelerators. Volume pricing, 15-day sample lead time.
Get Your Quote →