Edge AI Platforms

Accelerating AI-driven Inference for Edge Infrastructures

Volume 25.1 [Link]

1 [Link]
Bringing AI to the Edge

In today’s fast-paced digital landscape, Edge AI is truly transforming the way

network operations are conducted. By enabling real-time data processing
and analysis directly at the source, Edge AI significantly reduces latency,
minimizes bandwidth consumption, and enhances both privacy and security.
This revolutionary approach allows organizations to harness the power of
AI closer to where data is generated, leading to faster, more efficient, and
highly secure operations across a wide range of industries and applications.

At Lanner, we are at the forefront of this transformation, offering a

comprehensive suite of Edge AI platforms specifically designed to bring
advanced AI capabilities to the edge of your network. Our product
portfolio includes everything from compact Edge AI gateways to robust,
high-performance servers, all meticulously engineered for real-time
decision-making at the network’s edge. These solutions are built to meet
the demanding needs of applications such as computer vision, network
security, and telecommunications, where low latency, stringent security
measures, and efficient data handling are not just desirable, but essential.

By integrating Lanner’s Edge AI solutions, organizations can unlock new

levels of operational efficiency and security, ensuring they are well-
equipped to meet the challenges of today and tomorrow. Whether you
are looking to enhance network security, improve data processing speeds,
or implement cutting-edge AI applications, Lanner’s reliable and innovative
Edge AI platforms provide the foundation you need to succeed.

Jeans Tseng
CTO

[Link] 2
 Edge AI Platforms Overview
Integrated with AI hardware accelerators and pre-validated with AI solution partner, Lanner’s Edge AI
platforms enable edge computing solutions with specific requirements for low-latency, high-throughput,
and/or power efficiency in reliable and mission-critical applications.

These AI-accelerated hardware solution includes Edge AI servers, rugged computers, acceleration modules
and AI starter kits. Pre-validated by world’s leading software companies, Lanner provides system integrators
a turnkey, all-in-one edge AI platform, composed of a high-performing AI model and a powerful NVIDIA-
certified edge AI appliance.

Vision AI Network AI
Enable higher volume & low-latency video analytics Lanner provides comprehensive AI-accelerated network
solutions for mission-critical applications, such as smart security appliances designed to conduct proactive
retail, smart manufacturing, physical security, traffic security defense, enabling real-time threat detection,
monitoring, and more. adaptive responses, and predictive analytics.

Telco AI Agentic AI
Build scalable, GPU-accelerated multi-node edge Lanner provides MGX-based Edge AI servers by
computing platform for software-defined 5G/6G consolidating CPU, GPU, and DPU in a single unit,
applications, such as multi-access edge computing ensuring efficient handling of extensive AI models for
(MEC) and RAN intelligence. offline LLMs and generative AI deployment.

3 [Link]
Ecosystem Partners
Silicon Partners
NVIDIA
NVIDIA has been a pioneer in accelerated computing. The company’s invention of the GPU in 1999
sparked the growth of the PC gaming market, redefined computer graphics, ignited the era of
modern AI and is fueling the creation of the metaverse.

Intel
Intel OpenVINO™ toolkit is a comprehensive toolkit for quickly developing applications and solutions
that solve a variety of tasks, including emulation of human vision, automatic speech recognition,
natural language processing, and many others.

Computer Vision Software Partners

Chooch
Chooch is a leading Vision AI platform that combines Generative AI and Computer Vision to help
businesses automate repetitive, manual visual review tasks.

Data From Sky

Data From Sky offers TrafficXRoads, a highly optimized and fully configurable trajectory processing
engine that evaluates multiple detection tasks simultaneously within each camera’s field of view,
ensuring unparalleled accuracy and efficiency in traffic monitoring and management.

Ironyun
IronYun offers more than 30 advanced AI video analytics functions to bring intelligence and accuracy
to existing camera and video infrastructures.

Network Optix
Network Optix is at the forefront of video software innovation, developing sophisticated solutions
that convert video data into practical insights for businesses and industries of all kinds.

GoodVision
GoodVision provides automation tools in all stages of traffic projects, from AI traffic data collection
to traffic modelling and real-time traffic control.

Gorilla
Gorilla is a global solution provider in Security Intelligence, Network Intelligence, Business Intelligence
and IoT technology.

Milestone Systems
Milestone Systems aims to make the world see by empowering people, businesses and societies with
data-driven video technology.

Network Security & Telco Software Partners

F5
F5 partners with the world’s largest, most advanced organizations to optimize and secure every app
and API anywhere, including on-premises, in the cloud, or at the edge.

Arrcus Networks
The Arrcus Connected Edge (ACE) platform offers best-in-class networking with the most flexible
consumption model at the lowest total cost of ownership.

Namla
Namla offers a Software platform which allows B2B (Business To Business) and MSPs (Managed
Service Providers) customers to have a turnkey hybrid-cloud native Edge solution.

[Link] 4
 Vision AI

AI video analytics is the process of leveraging artificial intelligence and machine learning algorithms to
analyze and interpret video data. This allows for the automatic detection of objects, recognition of actions
and behaviors, and generation of business intelligence that can be used to improve safety, security, and
efficiency in a variety of applications.

Lanner offers a full range of AI-

accelerated Edge AI platforms that enable
video analytics solutions in industrial
settings. With embedded AI accelerators
to process deep learning inferencing in
real-time, these edge AI computers are
designed to handle multiple streams
of video simultaneously, detecting and
analyzing a wide range of objects, faces,
and actions in industrial environments.

Edge AI
Computer

Intel® Core™ Ultra Intel® Core™ Ultra Industrial Edge AI Industrial Edge AI

Model Name EAI-I500 EAI-I510 EAI-I730 EAI-I731

Intel® Core™ Ultra 7/5 Intel® Core™ Ultra 9/7/5

Processor System 12/13/14th Gen. Intel® Core i 12/13/14th Gen. Intel® Core i
(Meteor Lake-H/U) (Series 2)

NVIDIA L4, A2, A10, RTX Series NVIDIA A2, L4, RTX Series or
AI Acceleration Hailo-8 AI Accelerator Hailo-8 AI Accelerator
or any FHHL GPU cards any FHHL GPU cards
Support (by request) (by request)
(Up to 150W) (up to 75W)

Fanless Yes Yes YES YES

Max. Memory 96GB 96GB 96GB 96GB

2x 2.5” x 7mm (SATAIII) 1x M.2 M-key & 2x M.2 B-Key 1x M.2 M-key & 2x M.2 B-Key
Storage 2x 2.5” SATA SSD
1x M.2(NVMe) 2280 M-Key 4x 2.5” Removable HDD/SSD 2x 2.5” Removable HDD/SSD

Ethernet 3x 2.5GbE RJ45, 1x GbE RJ45 2x 2.5GbE RJ45, 2x 10GbE RJ45 2x 2.5GbE RJ45 2x 2.5 GbE RJ45

1x Console, 2x COM, 1x HDMI,

2x COM, 4x DIO, 4x USB, 2x COM, 8x USB 3.2, 1x HDMI,
I/O 2x COM, 4x DIO, 4x USB 1x DP 1.4, 8x USB, 4x DIO,
2x DP 1.2 2x Audio 1x DP 1.4, 4x DIO, Audio
Audio

1x M.2 B-key for 5G Sub6 / LTE 1x M.2 B-key for 5G Sub6 / LTE
1x M.2 B-key & E-key 1x 2230 M.2 E-key
Expansion 1x M.2 E-key for WiFi 1x M.2 E-key for WiFi
1x M.2 2280 M-key NVMe 1x Nano-SIM
2x PCIe*4, 2x PCIe*16 1x PCIe*16, 1x PCIe*4

Power +24VDC +24VDC 100 ~ 240VAC 12/24 VDC

Mechanical (W x D
287 x 180 x 76 mm 270 x 180 x 119 mm 374 x 419 x 250 mm 190 x 286 x 250 mm
x H)

Environment 0°C~40°C 0°C~40°C -25ºC~65ºC -25~65ºC

Win 10 IoT Win 10 IoT Win 10 IoT Win 10 IoT

Driver Support
Linux Dabian 11, Ubuntu 22.04 Linux Dabian 11, Ubuntu 22.04 Linux Dabian 11, Ubuntu 22.04 Linux Dabian 11, Ubuntu 22.04

CE/UKCA, Class A, UL/CB CE/UKCA, Class A, UL/CB

Certification CE/FCC Class A, UKCA, UL, CB CE/FCC Class A, UKCA, UL, CB
62368-1 62368-1

5 [Link]
Edge AI
Computer
NVIDIA Jetson Orin NVIDIA Jetson Orin NVIDIA Jetson AGX Orin NVIDIA Jetson AGX Orin

Model Name EAI-I131 EAI-I133 EAI-I233 EAI-I251

Processor System NVIDIA® Jetson Orin NX/Nano NVIDIA® Jetson Orin NX/Nano NVIDIA® Jetson AGX Orin NVIDIA® Jetson AGX Orin

AI Acceleration
NVIDIA Ampere Architecture NVIDIA Ampere Architecture NVIDIA Ampere Architecture NVIDIA Ampere Architecture
Support

Fanless YES YES 1x 12V Fan 1x 12V Fan

Max. Memory 16GB 16GB 64GB 64GB

1x M.2 M-key NVMe 1x M.2 M-key NVMe

Storage 1x M.2 M-key NVMe 1x M.2 M-key NVMe
eMMC 5.1 64GB eMMC 5.1 64GB

1x GbE RJ45, 1x 10GbE RJ45 1x GbE RJ45, 1x 10GbE RJ45

Ethernet 2x GbE RJ45 for PoE+ 3x GbE RJ45
1x 2.5GbE RJ45 2x 2.5GbE RJ45 w/ PoE+

2x COM, Audio, 2x USB 2.0 2x COM, 1x HDMI, 4x USB 3.2,

I/O 2x COM, 4x USB, 1x HDMI 1x COM, 1x HDMI, 4x USB 3.2
1x HDMI, 4x DIO 4x DIO, 8x GMSL2, Audio

1x M.2 B-key for 5G Sub6 / LTE 1x M.2 B-key for 5G Sub6 / LTE
1x M.2 B-key for 5G Sub6 / LTE 1x M.2 B-key for 5G Sub6 / LTE
Expansion 1x M.2 E-key for WiFi 1x M.2 E-key for WiFi
1x M.2 E-key for WiFi 1x M.2 E-key for WiFi
1x Nano-SIM 2x Nano-SIM

Power Typical +12VDC 12VDC Typical +12 or +24VDC 12/24VDC

Mechanical (W x D x H) 201 x 196 x 65 mm 167 x 179.5 x 30 mm 169.2 x 181.7 x 70 mm 201 x 196 x 95 mm

Environment -40°C~70°C 0°C~40°C 0°C~40°C -40°C~70°C

Linux Ubuntu 22.04, Linux Ubuntu 22.04, Linux Ubuntu 22.04, Linux Ubuntu 22.04,
Driver Support
NVIDIA Jetpack 6.2 NVIDIA Jetpack 6.2 NVIDIA Jetpack 6.2 NVIDIA Jetpack 6.2

Certification CE/FCC Class A, UL CE/FCC Class A, UL CE/FCC Class A, UL+CB CE/FCC Class A, UKCA, UL+CB

Edge AI
Computer
Vehicle Edge AI Vehicle Edge AI Railway Edge AI

Model Name EAI-V151 EAI-V330 EAI-R530

Processor System NVIDIA® Jetson Orin NX/Nano Intel® Atom® x6000 Series 13th Gen. Intel® Core i7-1300 Series

AI Acceleration
NVIDIA Ampere Architecture M.2 Hailo-8 AI Accelerator NVIDIA MXM A2000 GPU cards
Support

Fanless YES YES YES

Max. Memory 16GB 32GB 64GB

1x eMMC 128GB onboard 2x M.2 M-key NVMe

Storage 1x M.2 M-key NVMe
1x 2.5” SATA drive bay 2x 2.5” SATA drive bay in a caddy

Ethernet 1x GbE RJ45, 2x GbE PoE+ 6x GbE RJ45(4x GbE PoE) 4x 2.5GbE RJ45, 2x 2.5GbE PoE+

2x COM, 2x USB 3.2, 1x HDMI, 4x USB, 2x COM, 1x CAN 2.0

I/O 1x Console, 3x USB, 4 GNSS, 1x OOB
4x DIO, Audio, GNSS/G-sensor 4x DI, 2x DO, 2x HDMI, GNSS/G-sensor

2x M.2 M-key for 5G Sub6 / LTE

1x M.2 B-key for 4G LTE/ 5G sub6 1x M.2 B-key for 4G LTE/ 5G sub6 1x M.2 E-key for WiFi
Expansion
1x M.2 E-key for WiFi 1x M.2 E-key for WiFi 1x PCIe*8 Gen 3 MXM (optional)
4x PGN Modules (optional)

Power 12/24VDC 12 ~ 48VDC 24 ~ 110VDC

Mechanical (W x D x H) 201 x 196 x 65 mm 273.8 x 185 x 98 mm 438 x 111.25 x 300 mm

Environment -40ºC~70ºC -40ºC~70ºC -40ºC~70ºC

Linux Ubuntu 22.04, Linux kernel 2.6.X, Ubuntu 22.04,

Driver Support Linux Debian 10, Win 11 IoT
NVIDIA Jetpack 6.2 Debian 10, Win 10/11 IoT

CE/FCC Class A, UKCA, UL, CB, E-Mark CE/FCC Class A, UKCA, E13, UL/cUL & CB, CE/FCC Class A, MIL-STD-810G,
Certification
(E24) MIL-STD-810G EN50155, EN45545-2, UL+CB

[Link] 6
 Network AI

In the rapidly evolving landscape of network security, the integration of AI has emerged as a game-changer.
By employing advanced algorithms and machine learning, AI-powered network security has ushered in a new
era of proactive defense, enabling real-time threat detection, adaptive responses, and predictive analytics.

Specializing in providing network security hardware solutions, Lanner stands as a vanguard pioneering the
fusion of advanced network appliances and AI accelerators that empower AI-driven network security with
unparalleled efficiency, performance, and scalability. Lanner’s AI-powered network appliances can process
colossal volumes of data, enabling AI algorithms to swiftly discern anomalies, predict threats, and fortify
defences in real time.

AI Starter Kit: AI-Powered Next-Generation Firewall

Next-generation firewalls (NGFWs) are evolving to combat new cybersecurity threats such as ransomware,
zero-day attacks, and phishing, as well as ongoing threats like distributed denial of service (DDoS) attacks.
This evolution incorporates artificial intelligence and machine learning (AI/ML) to create a more dynamic
rules database. AI-driven detection capabilities enable NGFWs to quickly recognize and respond to cyber
threats by being trained on extensive databases of known threats, allowing them to infer and counter
new variants automatically. Lanner’s NCA-4240, featuring 14th Gen Intel® Core™ desktop processors,
is optimized for AI-based NGFW performance, balancing cost, power consumption, and form factor for
branch office applications.

7 [Link]
Network AI
Appliance
Intel Amston Lake Intel Raptor Lake AMD Genoa/Bergamo

Model Name NCA-1250 NCA-4240 NCA-5330

Intel® Atom x7000 & N Series x7425E/N97 12/13/14th Gen Intel® Core™ / Pentium®
Processor System AMD EPYC 9004 Series (Genoa/Bergamo)
or x7405C/x7835RE / Celeron®

AI Acceleration NVIDIA L4,

Integrated Intel® UHD Graphics 770 Integrated Intel® UHD Graphics 770
Support or any HHHL GPU Card (up to 75W)

Fanless Yes 3x Smart Fans 5x Individual Hot-swappable Cooling Fans

Memory 16GB 64GB 512GB

2x 2.5” HDD/SSD
1x M.2 (SATA) 2280 SKU A: 1x M.2 2242 SATA 2x 2.5” SSD/HDD
Storage
1x eMMC 16GB Onboard (By SKU) SKU B: 1x M.2 2242 SATA 1x M.2 2280 (SATAIII / PCIe*5)
& 1x M.2 2280 NVME

Ethernet 5x 2.5GbE RJ45, 1x 2.5GbE RJ45 (by SKU) 1x GbE RJ45, 8x 2.5GbE RJ45, 1x NIC Slot 1x GbE RJ45, x 4x NIC Slots

I/O 1x Console, 1x USB 3.0 1x Console, 2x USB 3.0 1x Console, 2x USB 3.0, 1x LOM (optional)

1x PCIe*8 Gen4 FH/HL (By SKU)

Expansion 1x M.2 3042/3050/3052 for 5G/LTE 1x PCIe*8 HH/HL (optional)
1x M.2 2230 E-key (By SKU)

Power 100 ~ 240VAC 100 ~ 240VAC 100 ~ 240VAC

Mechanical (W x D x H) 231 x 200 x 44 mm 438 x 321 x 44 mm 438 x 650 x 44 mm

Environment 0ºC-40ºC 0ºC-40ºC 0ºC-40ºC

Driver Support Linux Linux Linux

Certification RoHS, CE/FCC Class B RoHS, CE/FCC Class A, UKCA, UL RoHS, CE/FCC Class A, UL

Network AI
Appliance
Intel® Sierra Forest / Granite Rapids Intel® Sierra Forest / Granite Rapids Intel Emerald Rapids

Model Name NCA-6050 NCA-6250 NCA-6530

Intel® Xeon® 6 Family Processor Intel® Xeon® 6 Processors

Processor System 5th Gen Intel® Xeon® Scalable Processor
(Sierra Forest-SP/Granite Rapids- SP) with P cores & E cores

AI Acceleration NVIDIA H200, B40 or any FHFL NVIDIA L40, L4 NVIDIA L40S
Support GPU cards (up to 600W) or any FHFL GPU cards (up to 300W) or any FHFL GPU cards (up to 350W)

Fanless 4x Smart Fans 4x Smart Fans 6x Smart Fans

Memory 1024GB 1536GB 1536GB

2x 2.5’’ U.2 NVME SSD Hot-swappable

2x 2.5” Swappable HDD/SSD or 12x 2.5”
1x M.2 22110/2280/2242 M-key NVME
Storage 2x U.2 NVME SSD KIT NVMe Swappable HDD/SSD (by SKU)
1x M.2 2280/2242 M-key NVME
2x M.2 2280 NVME, 1x M.2 2280 SATA
1x M.2 2242/3042 SATA*4

Ethernet 1x GbE RJ45 2x GbE RJ45, 8x NIC Slots 2x GbE RJ45, 8x NIC Slots

I/O 1x Console, 2x USB 3.0, 1x LOM, miniDP 1x Console, 1x LOM, 2x USB 3.0, 1x VGA 1x Console, 1x LOM, 2 x USB 3.0, 1x VGA

Rear: 2x FHHL Single-Deck PCIe Cards

(Optional) 2x PCIe*16 FHFL or
Expansion 2x PCIe*16 Gen5
Front:2x FHFL Double-Deck PCIe Cards 2x PCIe*16 FHHL
(Optional)

Power 200~240VAC 100~240VAC 200~240VAC

Mechanical (W x D x H) 438 x 650 x 88 mm 438 x 650 x 88 mm 438 x 760 x 88 mm

Environment 0ºC-40ºC 0°C~40°C 0°C~40°C

Driver Support Linux Linux Linux

Certification CE/FCC Class A, UL CE/FCC Class A, UL, UKCA, VCCI CE/FCC Class A, UL

[Link] 8
 Telco AI & Agentic AI

AI RAN
Consolidating AI at the edge with telco networks through an edge AI platform accelerates
digital transformation and enhances value creation. Edge AI optimizes network operations by
analyzing massive volumes of network data in real-time, and empowers the RAN Intelligent
Controller (RIC), advancing RAN capabilities with AI to improve spectral and operational efficiency.

Lanner provides scalable, carrier-grade edge AI servers that consolidate CPU, GPU, and DPU resources into a
single appliance. These edge AI servers integrate Intel Xeon or NVIDIA GH200 processors, support multiple
PCIe*16 slots, and are compatible with NVIDIA L4 and L40S GPUs, NVIDIA BlueField-3 DPUs, and NVIDIA
ConnectX-7 network adapters.

Mainstream LLMs & GenAI

Edge AI serves as a crucial enabler for Generative AI and large language models (LLMs) by providing the
infrastructure to train and deploy these models locally. This approach enhances data privacy, reduces latency,
and eliminates the need for constant cloud communication, leading to faster, more secure, and cost-effective
AI solutions. By processing data at the edge, enterprises can harness the full potential of Gen AI and LLMs
while minimizing cybersecurity risks and transmission costs.

Lanner provides edge AI servers to enable enterprises to run large-scale LLMs and perform generative AI LLM
training and inferencing, delivering high-performance, low-latency AI processing directly at the edge.

9 [Link]
NVIDIA MGX Server: ECA-6051

Edge AI
Server
Intel Emerald Rapids Intel Emerald Rapids Intel® Sierra Forest Intel® Sierra Forest
/ Granite Rapids / Granite Rapids

Model Name ECA-5540 ECA-6040 ECA-6050 ECA-6051

5th/4th Gen Intel® Xeon® 5th/4th Gen Intel® Xeon® Intel® Xeon® 6 Family Processor Intel® Xeon® 6 Family Processor
Processor
Scalable Processor Scalable Processor (Sierra Forest-SP/Granite Rap- (Sierra Forest-SP/Granite Rap-
System
(Sapphire Rapids-SP/EMR-SP) (Sapphire Rapids-SP/EE/EMR-SP) ids-SP) ids-SP)

NVIDIA L4, L40S, B40 NVIDIA L4, L40S, B40 NVIDIA L4, L40S, B40
AI Acceleration NVIDIA L40S, H200, B40 or any
or any FHFL GPU cards or any FHFL GPU cards or any FHFL GPU cards (up to
Support FHFL GPU cards (up to 600W)
(up to 350W) (up to 350W) 350W)

Fanless 6x Smart Fans 6x Smart Fans 6x Smart Fans 6x Smart Fans

Memory 1024GB 1024GB 1024GB 1024GB

4x 2.5” HDD/SSD or 4x U.2

2x 2.5” HDD/SSD or 2x U.2 4x E1.S SSD 2x E1.S SSD
Storage 1x M.2 NVME M-key
2x M.2 M-key NVME 2x M.2 NVMe M-key 2x M.2 NVMe (PCIe)
1x M.2 2280 M-key

Ethernet 1x GbE RJ45 for MGMT 1x GbE RJ45 for MGMT 1x GbE RJ45 for MGMT 1x GbE RJ45 for MGMT

1x Console, 2x USB 3.0, 1x LOM, 1x Console, 2x USB 3.0, 1x LOM, 1x Console, 2x USB 3.0, 1x LOM, 1x Console, 1x USB 3.0, 1x LOM,
I/O
miniDP miniDP miniDP miniDP

1x PCIe*16 FHFL (by CPU)

2x PCIe*16 FHFL, 2x PCIe*8 LP 4x PCIe*16 FHFL
Expansion 2x PCIe*16 LP or 1x PCIex8 3x PCIe*16 FHFL
1x OCP3.0 Slot 1x PCIe*16 FHHL
FHHL, 1x OCP3.0 Slot

Power 110~240VAC 200~240VAC 200~240VAC 200~240VAC

Mechanical 438 x 580 x 44 mm 438 x 580.1 x 88 mm 438 x 760 x 88 mm 438 x 420 x 88 mm

Environment 0ºC-40ºC 0ºC-40ºC 0ºC-40ºC 0ºC-40ºC

Driver Support Linux Linux Linux Linux

RoHS, CE/FCC Class A, UL,

Certification CE/FCC, Class A, UL CE/FCC Class A, UL CE/FCC Class A, UL
NEBS (Level 3)

[Link] 10
 AI Starter Kit

Lanner’s Edge AI Starter Kits aim to deliver an all-in-one platform composed of AI-accelerated hardware
and application-tailored software to accelerate time-to-market AI deployment at the industrial edge. This
comprehensive kit is designed for various applications, including vision inspection, predictive maintenance,
and operator monitoring.

AGV/AMR Traffic Monitoring Robotics Speech AI Medical Imaging

Applications

Machine Learning Deep Learning Natural Language Processing Computer Vision

Modules

SDK/Library

Operating
System

Edge AI Platform

NVIDIA Software Supported

Intelligent Telecommunication
GenAI & LLMs Cyber Security
Video Analytics & Networking

NVIDIA AI Enterprise NVIDIA Metropolis NVIDIA Morpheus NVIDIA Aerial/DOCA

SIEM/SOAR
Reasoning Speech & Translation BioMedical Content Generation Route Planning
Preprocess Inference Post Process Take Action

App Logs

NVIDIA Morpheus
Community Models NVIDIA Models Custom Models

Cloud Logs

NIM Microservices CUDA-X Microservices RAPIDS TensorRT

Security Advisory Enterprise Support Cluster Management Infra Optimization Production Branch

NVIDIA BlueField
Accelerated Servers
NVIDIA AI Enterprise

NVIDIA SmartNIC / DPU NVIDIA GPU

Network and Infrastructure Acceleration Application Acceleration
Cloud | Data Center | Workstations | Edge
Converged Card

EAI-I131/130 EAI-I133

ECA-5540 ECA-5540
ECA-6040

EAI-I730 EAI-I731
ECA-6051 NCA-6250 ECA-6040

LEC-2290 NCA-6530/6520 ECA-6051

11 [Link]
 NVIDIA
Lanner provides Edge AI appliances that support the NVIDIA JetPack SDK, offering robust
tools for AI development and deployment. These appliances leverage NVIDIA’s powerful
AI capabilities, enabling seamless integration and efficient execution of complex AI
models at the edge..

Intel
Powered by the Intel® Distribution of OpenVINO™ toolkit, EIS accelerates development,
enabling quick integrations of pre-trained models (e.g., Tensorflow, Caffe, etc.) for
object recognition, classification, and facial recognition in vision-based solutions.

Edge AI
Starter
Kit
NVIDIA Jetson Orin NVIDIA Jetson Orin NVIDIA A2 GPU NVIDIA L4 GPU

Model Name EAI-I131 EAI-I133 LEC-2290E LEC-2290L

Processor System NVIDIA® Jetson Orin NX/Nano NVIDIA® Jetson Orin NX/Nano Intel® Core™ i7 Intel® Core™ i7

GPU NVIDIA® Jetson Orin NX/Nano NVIDIA® Jetson Orin NX/Nano NVIDIA® A2 60W NVIDIA® L4 72W

Memory LPDDR5 16/8/4 GB (By SKU) LPDDR5 16/8/4 GB (By SKU) 64GB DDR4 64GB DDR4

128GB mSATA 128GB mSATA

Storage M.2 128GB NVMe M.2 128GB NVMe
256GB SSD 256GB SSD

Linux Ubuntu 22.04, Linux Ubuntu 22.04,

Supported OS Ubuntu 22.04 Ubuntu 22.04
NVIDIA Jetpack 6.2 NVIDIA Jetpack 6.2

Intel OpenVINO Intel OpenVINO

Supported SDK NVIDIA AI Stacks NVIDIA AI Stacks
NVIDIA AI Stacks NVIDIA AI Stacks

Edge AI
Starter
Kit
NVIDIA A10 GPU NVIDIA A2 GPU NVIDIA L4 GPU Intel UHD Graphics 770

Model Name EAI-I730M EAI-I731E EAI-I731L NCA-4240C

Intel®14th Gen. Core i Proces- Intel®14th Gen. Core i Proces- Intel®14th Gen. Core i Proces- Intel®14th Gen. Core i Proces-
Processor System
sors with R680E Chipset sors with R680E Chipset sors with R680E Chipset sors with Q670E Chipset

Intel® UHD Graphics 770 GPU.

GPU NVIDIA® A10 150W NVIDIA® A2 60W NVIDIA® L4 72W
1x PCIe*8 Gen4 FHHL

2x DDR4 3200MHz SODIMM Up 2x DDR4 3200MHz SODIMM Up 2x DDR4 3200MHz SODIMM Up 2x DDR5 5600MHz UDIMM Up
Memory
to 64GB to 64GB to 64GB to 32GB

1x M.2 B-key SATA 1x M.2 B-key SATA 1x M.2 B-key SATA 2x 2.5” HDD/SSD
Storage 1x M.2 M-key NVMe 1x M.2 M-key NVMe 1x M.2 M-key NVMe 1x M.2 E-key NVME (PCIe
4x 2.5” SATA drive bay 2x 2.5” SATA drive bay 2x 2.5” SATA drive bay Gen4*4)

Supported OS Ubuntu 22.04 Ubuntu 22.04 Ubuntu 22.04 Ubuntu 22.04

Intel OpenVINO Intel OpenVINO Intel OpenVINO Intel OpenVINO

Supported SDK
NVIDIA AI Stacks NVIDIA AI Stacks NVIDIA AI Stacks NVIDIA AI Stacks

[Link] 12
 AI Acceleration Cards

Lanner’s Falcon PCIe AI Acceleration Cards offer a convenient solution for engineers seeking to offload CPU
load for low-latency deep learning inference. Equipped with high-density AI processors, the Falcon supports
1 to 6 Hailo-8™ AI processors, providing a modular Edge AI solution with exceptional processing power and
energy efficiency

The Falcon AI Accelerator Cards enable legacy devices, including NVRs, Edge AI boxes, industrial PCs, and
robots, to run video-intensive, mission-critical Edge AI applications via a standard PCIe interface. These
applications encompass video analytics, traffic management, access control, and more.

Falcon H8 Falcon Lite

Model No. AI VPU Remark Model No. AI VPU Remark

FALCON-H8A 6x Hailo-8™ AI Processor Commercial-grade Falcon H8L16A 2x Hailo-8™ AI Processor Commercial-grade

Falcon H8L16D 4x Hailo-8™ AI Processor Commercial-grade

FALCON-H8B 5x Hailo-8™ AI Processor Commercial-grade
Falcon H8L16E 2x Hailo-8™ AI Processor Industrial-grade
FALCON-H8C 4x Hailo-8™ AI Processor Commercial-grade
Falcon H8L16H 4x Hailo-8™ AI Processor Industrial-grade
FALCON-H8D 6x Hailo-8™ AI Processor Industrial-grade
Falcon H8L8A 2x Hailo-8™ AI Processor Commercial-grade

FALCON-H8E 5x Hailo-8™ AI Processor Industrial-grade Falcon H8L8D 4x Hailo-8™ AI Processor Commercial-grade

FALCON-H8F 4x Hailo-8™ AI Processor Industrial-grade Falcon H8L8EE 2x Hailo-8™ AI Processor Industrial-grade

Falcon H8L8H 4x Hailo-8™ AI Processor Industrial-grade

Falcon H8L4A 1x Hailo-8™ AI Processor Commercial-grade

Falcon H8L4B 2x Hailo-8™ AI Processor Commercial-grade

Falcon H8L4E 1x Hailo-8™ AI Processor Industrial-grade

Falcon H8L4F 2x Hailo-8™ AI Processor Industrial-grade

13 [Link]
 NVIDIA GPU Supported
Lanner’s Edge AI platforms support a wide range of NVIDIA GPU cards, delivering scalable and powerful AI
performance at the edge. These platforms enable real-time data processing, deep learning, and AI-driven
analytics across various applications, from smart cities to industrial automation. With Lanner, businesses can
deploy and easily scale AI solutions to meet growing demands.

Model Name L4 L40S B40 H200

ECA-4035 ✔ ✔

ECA-5540 ✔ ✔ ✔

ECA-6040 ✔ ✔ ✔

ECA-6050 ✔ ✔ ✔ ✔

ECA-6051 ✔ ✔ ✔

NCA-6050 ✔ ✔ ✔ ✔

NCA-5330 ✔

RTX 2000E RTX 4000 SFF RTX 4000E RTX 5000E RTX 6000E
Model Name A2 L4 A10
Ada Ada Ada Ada Ada

LEC-2290 ✔ ✔

EAI-I731 ✔ ✔ ✔ ✔

EAI-I730 ✔ ✔ ✔ ✔ ✔ ✔ ✔ ✔

[Link] 14
Corporate Taiwan

Lanner Electronics Inc. 立端科技股份有限公司

7F, No.173, Sec.2, Datong Rd. 221新北市汐止區
Xizhi District, 大同路二段173號7樓
New Taipei City 221, Taiwan T: +886-2-8692-6060
T: +886-2-8692-6060 F: +886-2-8692-6101
F: +886-2-8692-6101 E: contact@[Link]
E: contact@[Link]

USA China
Lanner USA 立华科技
47790 Westinghouse Drive 北京市昌平区
Fremont, CA 94539 回龙观回南北路果栋LOFT9层
T: +1-855-852-6637 T: +86 010-82795600
F: +1-510-979-0689 F: +86 010-62963250
E: sales_us@[Link] E: services@[Link]

Canada Europe
Lanner Canada Lanner Europe B.V.
6285 Northam Wilhelmina van Pruisenweg 104
Dr. Unit 112 Mississauga 2595 AN The Hague,
ON L4V 1X5 The Netherlands
T: +1 877-813-2132 T: +31 70-701-3256
F: +1 905-362-2369 E: sales_eu@[Link]
E: sales_ca@[Link]

Please verify specifications before quoting. All product specifications are subject to change without
notice. No part of this publication may be reproduced in any form or by any means, electronic,
photocopying or otherwise without prior written permission of Lanner Electronics Inc. All brand
names and product names are the trademarks or registered trademarks of their respective companies.

© Lanner Electronics Inc., 2025 [Link]