MPN: EC-ThorT5000
Firefly EC-ThorT5000
Jetson Thor T5000 · 2,070 FP4 TFLOPS · Industrial Edge AI

Firefly's first Jetson Thor T5000 industrial PC — 8× GMSL2 cameras, 4× 10GbE + 4× GbE PoE, dual NVMe, and up to 92-channel 1080p decode in a rugged all-aluminum chassis.

Specifications

AI Module

NVIDIA Jetson Thor T5000

AI Performance

2,070 TFLOPS FP4 (sparse)

CPU

14-core Arm Neoverse-V3AE @ up to 2.6 GHz

GPU

NVIDIA Blackwell — 2560 CUDA cores, 96 Tensor Cores, 10-TPC MIG

Memory

128GB 256-bit LPDDR5X @ 273 GB/s

Storage

2× M.2 M-Key 2280 (NVMe)

Video Output

2× HDMI 2.0 (up to 4Kp60)

Camera Input

8× GMSL2 via 2× FAKRA connectors

Ethernet

4× 10GbE RJ45 + 4× GbE RJ45 (PSE/PoE)

USB

4× USB 3.0 Type-A + 2× USB-C (debug + OTG)

Expansion

M.2 B-Key (5G) · Mini PCIe (4G LTE) · M.2 E-Key (Wi-Fi 6)

GNSS

Built-in GPS module

Video Decode

Up to 92× 1080p30 H.265 streams

Power

24V DC (9–36V) via 5.5×2.1mm barrel jack

Dimensions

277.95 × 136.09 × 88.0 mm

Operating Temp

−20°C ~ 60°C

OS Support

JetPack SDK / Ubuntu 24.04

Stock

Available

Price

Quote

Overview

The Firefly EC-ThorT5000 is an industrial edge AI PC built around NVIDIA's Jetson Thor T5000 module, delivering up to 2,070 TFLOPS FP4 of AI compute. One of the first commercial systems on the Jetson Thor platform, it reuses Firefly's proven EC-series rugged enclosure to bring Blackwell architecture to high-density edge workloads such as robotics, smart-city vision, and multi-camera AI inference.

At its core, the Jetson Thor T5000 combines a 14-core Arm Neoverse-V3AE CPU with a 2560-core Blackwell GPU (96 Tensor Cores, 10-TPC MIG) and 128GB of 256-bit LPDDR5X at 273 GB/s. Firefly pairs this with 8× GMSL2 camera inputs (via 2× FAKRA), 4× 10GbE + 4× GbE PoE Ethernet, and dual M.2 NVMe storage, plus expansion for 5G, 4G LTE, and Wi-Fi 6 with a built-in GPS.

The Blackwell video engine decodes up to 92 simultaneous 1080p30 H.265 streams (or 46× 1080p60), making the EC-ThorT5000 ideal for dense video analytics and smart-city deployments. Software support spans the full JetPack SDK on Ubuntu 24.04 with CUDA, TensorRT, and cuDNN, plus PyTorch, TensorFlow, PaddlePaddle, and Docker — enabling local LLM and image-generation workloads via Ollama and ComfyUI (Deepseek-R1, Llama, Qwen, Stable Diffusion).

The all-aluminum chassis runs on 9–36V DC input across a −20°C to 60°C range for 24/7 industrial operation, measuring 277.95 × 136.09 × 88.0 mm.

Need Firefly EC-ThorT5000?

Jetson Thor T5000 industrial edge AI PC with 8× GMSL2 and 4× 10GbE. Contact QS Compute for pricing, availability, and configuration.

Request Quote