NextServer AI 5G: The Portable AI 5G Base Station

Bring a Real-Time AI 5G Base Station With You Anywhere you need to be
Built for the Edge
Engineered for maximum performance-per-watt, the NextServer AI 5G is optimized for the Ubuntu 24.04 LTS ecosystem and Canonical Kubernetes, providing a stable, carrier-grade platform for mission-critical applications.
- Processor: AmpereOne® 192-core arm64-bit CPU for massive parallel processing.
- Memory: 8 channels of DDR5 memory for high-bandwidth data handling.
- Storage: Terabytes of high-speed NVMe storage to support data-intensive pipelines.
- AI Acceleration: Optional NVIDIA RTX 4000 SFF Ada for high-demand inference.
- DeepSig: Option for integration of DeepSig software stack for AI inference on Ampere CPU and/or NVIDIA GPU for Spectrum Intelligence: Leverages OmniSIG® for real-time RF situational awareness
- Connectivity: Built-in 5G radio unit (RU) with portable antenna and Starlink Mini backhaul support.
- Form Factor: TSA/IATA Compliant airline carry-on (22” x 14” x 9”) for global mobility.
- Easy Management: AI applications are quickly deployed and managed by Kasm Workspaces
The Strategic Advantage
Traditional setups for field operations—ranging from disaster response and conservation to border security and autonomous systems—require separate appliances for packet capture, analysis, and AI processing. The NextServer AI 5G consolidates this stack into a single, portable unit.
- Low-latency computing at the edge and in remote locations with limited or unreliable infrastructure
- Secure, resilient communications with private 5G deployments that ensure data never leaves your local network.
- Reliability via built on SRS/LF OCUDU open-source carrier-grade software.
- Rapid global deployability via airline carry-on and connection via satellite backhaul anywhere on the map.
Whitepaper
AI-Native Deployable 5G Systems and Autonomous Edge Orchestration
This white paper outlines the development and strategic value of the NextServer AI 5G Fly-Away Kit (FAK), a portable "Data Center in a Suitcase" designed for remote, tactical, and infrastructure-denied environments.
Powered by a 192-core AmpereOne ARM processor and optimized for open-source O-RAN software, the system functions as a "Tactical AI Cloud" capable of supporting 2,000 simultaneous users and massive data throughput.
It is engineered for tasks such as orchestrating autonomous drone swarms via the SteelEagle platform, performing real-time AI-driven wildlife or border surveillance, and providing advanced RF spectrum awareness using DeepSig OmniSIG.
By consolidating high-performance computing, massive NVMe storage (up to 1PB), and satellite backhaul into a ruggedized carry-on, the FAK eliminates vendor lock-in and provides a secure, TAA-compliant solution for mission-critical deployments.
Use Case: Wildlife Conservation AI 5G Hotspot

- The Mission: Real-time wildlife detection and species classification in remote environments.
- The AI Pipeline: Uses YOLO + MegaDetector and BearID facial recognition to identify and track individual animals via 5G-connected trail cameras and drones.
- The 5G Advantage: Enables researchers to use 5G-connected mobiles and tablets to catalog wildlife directly at the edge.
Wildlife AI tools and carrier-grade 5G software integrated into portable 5G AI hotspot empower research and conservation teams in the field with real-time wildlife detection and species classification in remote environments.
Field teams deploy 5G-capable trail cameras, autonomous drones, and mobile applications powered by a portable fly-away kit airline carry-on 5G base station combined with edge AI compute and Starlink Mini internet backhaul. Video and images captured by camera traps and drones are streamed to the 5G AI hotspot, where they are processed through AI pipeline performing animal detection, species classification, and bear facial recognition (for bear identification), timestamping, and geotagging. This enables accurate cataloging and tracking of wildlife directly at the edge.
This solution delivers secure, high-performance AI computing and connectivity in areas where traditional infrastructure is unavailable.
5G Dashboard built into OCODU open source RAN project https://ocudu.org/
Click to view
With the Wildlife Conservation AI 5G Hotspot, open-source OCUDU software, and a radio unit, private 5G becomes portable allowing you to connect to any phone, including the one in your pocket, and run AI-powered workloads at the edge.
Technology used (HW platform, OS, CaaS, 5GC, RU, UE, etc.)
AI Software:
- TRAPPER Wildlife Conservation AI (https://os-conservation.org/trapper/)
- Animal detection
- Species classification
- Geolocated and timestamped
- Conservation team tools used via 5G connected mobiles, tablets and laptops.
- BearID (https://bearresearch.org)
- Facial recognition for bears identification
- Ampere-Optimized AI Models (https://amperecomputing.com/solutions/ampere-ai)
- YOLO based MegaDetector for animal detection
- SpeciesNet for species classification
Hardware & Infrastructure Software:
- NextComputing NextServer AI 5G fly away kit (airline carry-on form factor)
- AmpereOne® 192 core arm64-bit processor, 8 channels DDR5 memory, TBs of NVMe storage, optional NVIDIA RTX 4000 SFF Ada
- Built-in 5G radio unit (RU) and includes portable cellular antenna
- Operating System & Container Platform
- Ubuntu 24.04.3 LTS
- Canonical Kubenetes (carrier-grade container orchestration)
- 5G Network Stack (https://ocudu.org)
- Private 5G deployment
- SRS developed LF OCUDU open source carrier grade 5G software
Mobile World Congress 2026
From March 2-5 in Barcelona, Canonical is presenting a working wildlife conservation platform that combines open source 5G, AI, and cloud-native infrastructure in a travel friendly form factor.
At the center is a portable 5G AI hotspot built on Ubuntu and Canonical Kubernetes, running on Arm-based Ampere servers. It connects drones, trail cameras, and researcher applications over a built-in private 5G network, with optional satellite backhaul.
Its purpose is to process wildlife data including photos and videos where it is collected, extract insights with AI, and share structured results globally.
A special thank you to Karolina Kuczkowska from the Open Science Conservation Fund for this video presentation.
This is an excellent look at the AI 5G station configuration for trail camera analysis and how it creates reliable access in areas lacking communication infrastructure, where conservation work matters the most.
One Platform, Infinite Missions

First Responder & Disaster Response Teams
- The Mission: Autonomous site operations and predictive actions in disaster zones.
- The AI Pipeline: Utilize sensor data to detect possible hazards.
- The 5G Advantage: Provides a private 5G bubble to maintain communication independent of unstable networks

Drone Swarms & Orchestration
- The Mission: Collaborative perception and ground-control for autonomous swarm intelligence.
- The AI Pipeline: Acts as a centralized "brain" to offload complex mapping and trajectory planning from lightweight drones.
- The 5G Advantage: Orchestrates high-speed telemetry and real-time video streaming between multiple flight units and the FAK.

Precision Agriculture
- The Mission: Nomadic farm management and autonomous crop health analysis.
- The AI Pipeline: Processes multi-spectral drone imagery to identify pest infestations or crop disease at the edge.
- The 5G Advantage: Coordinates swarms of weeding and irrigation robots across thousands of acres where cellular towers are non-existent.

Cybersecurity & Threat Hunting
- Application: OT/SCADA protection and network forensics in remote locations. Example: Ingesting host logs and network traffic through Packet Continuum, Suricata, and Zeek to identify threats in real-time at remote industrial sites
System Specs
| 5G Radio and OCUDU |
Integrated 5G radio unit (RU), portable cellular antenna. SRS LF OCUDU carrier grade 5G software stack. Optional Ampere AI stack APIs and use case specific development/integration services from NextComputing. |
|---|---|
| CPU |
|
| Memory |
Up to 2TB RAM |
| PCI Expansion |
(7) PCI Express slots for ½ height PCI Express cards to ¾ length cards or 9” length |
| Storage |
|
| Sustained Read/Write |
NVME to 15GB/second per drive (up to 8 drives total). Up to 1PB NVMe storage. |
| Network |
|
| AI |
AmpereOne processor for AI Compute. Option for two AmpereOne processors w/o NVIDIA GPU. Optional NVIDIA RTX 4000 SFF Ada , RTX 2000 Ada, and L4 Tensor Core GPU(s) |
| Security | Secure Boot UEFI compliant BIOS, Boot Guard, TPM2.0 |
| Remote Management | IPMI, HTML5 and API based out of band management with MAC addresses identification, one time boot on next reboot option, PXE boot option, Redfish compatible Out of Band (OBB) management connection |
| Available Operating Systems |
|
| Power |
|
| Environmental |
|
| Physical |
9” H x 21.75” W x 13.875” D (complete system including operational hard case) |
| Warranty | 1 year parts and labor. 2nd and 3rd year warranty options |
| Ease of Service and Upgrade | Most storage options are removable without tools. Easy access for install and removal of PCI Express cards. Easy to service other system elements. Modular design |
| Purpose-Built Solutions | NextComputing designs/manufactures its NextServer-X product line in the US (TAA compliant) and offers custom configurations, chassis modifications (https://solutions.nextcomputing.com/product-customization/), software integration services (https://solutions.nextcomputing.com/software-system-development-integration/), and (https://solutions.nextcomputing.com/services/software-services/), other FAK kitting options (https://solutions.nextcomputing.com/fly-away-kits/), and Private label branding options (https://solutions.nextcomputing.com/branding/). Includes Purpose built application integration and configuration control management services (https://solutions.nextcomputing.com/configuration-management/) |











