PE3000N

Next-generation Edge AI platform accelerated by NVIDIA^® Jetson Thor™, 5 GbE, GMSL, supporting LTE/5G/GNSS module, DIO, COM, 12-60V

Next-Gen Edge AI Supercomputer with NVIDIA Jetson T5000™
Rugged, Low-profile and scalable chassis design
Optional second stack enables project-specific I/O for vertical (CAN, QSFP28, POE) expansion, within 2U height
DC 12 to 60V wide range of DC or Battery power input with Ignition
Supports multiple generative AI pipelines and sensor fusion for next-gen robotics and automation
Full support for NVIDIA AI stacks (NVIDIA Isaac™, Metropolis, Holoscan) for fast cloud-to-edge deployment
Subscribe for updates here

Compare

PE3000N Jetson Thor

PE3000N Datasheet

ASUS IoT PE3000N is a compact edge-AI supercomputer designed specifically for the demands of next-generation robotics. Accelerated by NVIDIA^® Jetson Thor™ module: NVIDIA Jetson T5000™, it delivers an impressive 2,070 FP4 TFLOPS through a Blackwell GPU, a 14-core Arm CPU, and 128GB of LPDDR5X memory. This powerful combination enables PE3000N to efficiently run advanced generative VLM/LLM models, real-time intelligent video analytics, and fully autonomous control applications. Despite its small form factor, PE3000N offers robust performance, making it ideal for integration into space-constrained robotic systems without sacrificing computing power.

Built for versatility and adaptability, PE3000N features a scalable chassis with modular I/O layers supporting PoE, GMSL, CAN, QSFP28, and more, allowing seamless customization to meet evolving sensor and network demands. Its wide 12–60V input with ignition support ensures reliable, battery-friendly operation across diverse environments. Additional features include PTP/PPS for precise sensor synchronization, on-board TPM 2.0 for enhanced data security, and MIL-STD-810H military-grade rugged connectors that support up to four optional 25GbE links and 16 GMSL cameras for high-bandwidth sensor fusion. From factory floors and off-road vehicles to smart-city infrastructure, PE3000N delivers cloud-class intelligence at the extreme edge.

ASUS IoT PE3000N Edge AI Computer key visual banner featuring a sleek, black device with ports. Text highlights Nvidia Jetson AGX Thor power and AI capabilities.

Accelerated by NVIDIA^® Jetson Thor™

Rugged. Modular.
Accelerating Physical AI.

ASUS IoT PE3000N Edge AI Computer

2,070 FP4 TFLOPS Rugged Reliability
Flexible Expansion Compact Design

The Ultimate Platform for Humanoid Robotics

ASUS IoT PE3000N is a compact edge-AI platform accelerated by NVIDIA^® Jetson Thor™, delivering up to 2,070 FP4 TFLOPS – 7.5X the performance and 3.5X the efficiency of previous platforms. Its Blackwell GPU, 14-core Arm CPU and 128GB LPDDR5X run cutting-edge generative VLM/LLM, real-time video analytics and autonomous-robotics workloads, making it ideal for deployment across factory, logistics, and smart city installations – bringing cloud-level intelligence to the extreme edge.

Unmatched AI Performance

A futuristic robot with a sleek blue and white design stands behind a compact ASUS IoT PE3000N Edge AI computer. The scene conveys innovation and advanced technology.

Up to

2,070

FP4 TFLOPS*
7.5X

higher AI compute
than NVIDIA Jetson AGX Orin™
3.5X

better power efficiency
than NVIDIA Jetson AGX Orin™

*Compared with NVIDIA Jetson AGX Orin™

Modular Layer for Flexible Deployment

ASUS IoT PE3000N’s compact core system and modular chassis lets you stack plug-and-play I/O layers – PoE, GMSL, QSFP28, CAN, extra NVMe and more – tailoring or upgrading the system without touching the core.

Humanoid Robots
Factory Automation
Intelligent Video Analytics

Rich I/O for Expandability

Besides the core system, PE3000N offers multiple stack options for project-specific I/O for vertical expansion via CAN, QSFP28, and POE – all within 2U height.

1 2 3 4 5 6 7 8 9

1Power Button

2USB 3.2 Gen 2 Type-C

3USB 3.2 Gen 2 Type-C OTG

44 x USB 3.2 Gen 2 Type-A

5HDMI 2.0

61 GbE LAN

75 GbE LAN

8Terminal Block DC-in

9Audio In/Out

4 x 10G M12, GMSL Stack I/O Mechanical Layout.

1 2 3 4 5 6 7 8 9 10 11

14 x M12 10 GbE POE* LAN
*POE optional

2Power Button

3USB 3.2 Gen 2 Type-C

4USB 3.2 Gen 2 Type-C OTG

54 x USB 3.2 Gen 2 Type-A

6HDMI 2.0

71 GbE LAN

85 GbE LAN

9Terminal Block DC-in

10Audio In/Out

118 x GMSL2 mini FAKRA

6 x LAN port Stack I/O Mechanical Layout.

1 2 3 4 5 6 7 8 9 10

16 x 1 GbE POE* LAN
*POE optional

2Power Button

3USB 3.2 Gen 2 Type-C

4USB 3.2 Gen 2 Type-C OTG

54 x USB 3.2 Gen 2 Type-A

6HDMI 2.0

71 GbE LAN

85 GbE LAN

9Terminal Block DC-in

10Audio In/Out

Customization Stack I/O Mechanical Layout.

1 2 3 4 5 6 7 8 9 10

1Configured and Customized for Vertical Focus

2Power Button

3USB 3.2 Gen 2 Type-C

4USB 3.2 Gen 2 Type-C OTG

54 x USB 3.2 Gen 2 Type-A

6HDMI 2.0

71 GbE LAN

85 GbE LAN

9Terminal Block DC-in

10Audio In/Out

Engineered for the Extreme

PE3000N is designed for real-time reasoning in general robotics and physical AI. Its 12 to 60V wide-range DC input with ignition control ensures reliable power delivery, making PE3000N the ideal solution for mission-critical deployments. PTP/PPS support enables precise sensor synchronization, while it delivers unwavering performance in extreme temperatures ranging from -20°C to +60°C. The system also boasts MIL-STD-810H certification for exceptional ruggedness, and on-board TPM 2.0 provides enhanced data security.

12-60V Range DC-in
Sensor Fusion by PTP/PPS Technology
Ignition power control
‑20°C to +60°C temperatures
IEC 62443-4-1
MIL-STD-810H compliant

Versatile industrial AI applications

A machine operates on a conveyor belt, processing items alongside another machine in a factory setting.

Machine vision

High-performance GPUs support deep-learning image recognition and real-time inspection
PCIe 5.0 for frame grabbers and motion-control integration
Rich I/O: USB, PoE, and isolated DIO for flexible camera and sensor connections
MIL-STD-810H-certified ruggedness for factory-floor deployment