Prev
Next
AIP-FR68S
• Intel 12/13/14th Gen Core™ i9/i7/i5 processor, TDP up to 65W
• PCIe Gen5 x16 Support: Enables seamless integration with the latest GeForce RTX-5090 GPU
• Integrated PCIe Gen5 Switch: Supports unified computing across multiple Qualcomm AI100 Ultra cards
• Swappable 1600W Power Supply: Delivers ample, replaceable power for high-performance GPUs and AI accelerators
• Built-In GPU Holder: Ensures GPU stability, even in rugged operating environments
Overview
Aetina MegaEdge AIP-FR68S - Scalable Edge AI Workstation for GenAI Workloads
The AIP-FR68S edge AI workstation is purpose-built for on-prem AI, LLM and advanced AI vision inference, supporting both data center and workstation GPUs—including NVIDIA RTX PRO 6000 Blackwell /6000 Ada and up to three Qualcomm Cloud AI 100 Ultra accelerators (870 TOPS INT8 each).
Featuring a modular chassis design, 1600W PSU, and PCIe Gen5 switch, AIP-FR68S offers scalable GPU flexibility and high-performance edge computing. Integrated with the Qualcomm AI Inference Suite, it enables real-time LLM fine-tuning, multimodal analysis, and autonomous command generation—empowering AI agents, on-prem inference, and advanced visualization. Paired with NVIDIA AI Workbench, it accelerates the development of fine-tuned generative AI models for diverse applications such as AOI-based defect detection, image enhancement in smart manufacturing, contextual knowledge management, and semantic querying for agentic AI.
On-Prem Gen AI Performance
192GB DDR5 Memory
Next-Gen AI Core
Streamline AI Development
| ● Maximize AI Performance with Ultra-Fast 192GB Memory |
| Supports up to 192GB of DDR5 5600MT/s via four 48GB U-DIMM slots—delivering the bandwidth and capacity AI workloads demand. |
| ● Powerful GPU and AI Accelerator Support for High AI Computing Power |
| Features a PCIe Gen5 high-speed switches, enabling models twice as large (up to 200 billion parameters) or twice the inference performance, enhancing accuracy in natural language processing, image recognition, and predictive analytics applications requiring context-rich and complex AI models. The AIP-FR68S supports up to three Qualcomm Cloud AI 100 Ultra cards and is compatible with NVIDIA RTX GPUs and latest NVIDIA Blackwell architecture GPUs. |
| ● Streamlined AI Development for Flexible Deployment |
| AIP-FR68S turnkey solution supports NVIDIA AI Workbench and Qualcomm® AI Inference Suite, empowering developers to build, test, and scale large language models for on-premises or cloud deployment. Enables comprehensive generative AI applications in air-gapped environments, including voice agents, image generation, RAG, and code generation. |
| ● Enterprise-level artificial intelligence innovation |
| AIP-FR68S integrates AI visual detection technology to provide AI solutions suitable for enterprises for object picking, defect detection, image enhancement, and currently the most popular LLM (Large Language Model) applications. |
| ● Reliable Remote Management with EdgeEye |
| Seamlessly manage edge systems with Aetina's EdgeEye platform. Includes validated out-of-band module for remote recovery, ensuring rapid and reliable system maintenance. |
| ● RoHS Compliant for a Greener Future |
| System-wide design aligns with RoHS regulations and ESG sustainability goals. |
Innovative Gen AI Engine for Every Enterprise Need
Unlock on-prem generative AI, advanced AI vision with the AIP-FR68S—supporting NVIDIA RTX Pro 6000 Blackwell/ 6000 Ada and Qualcomm Cloud AI 100 Ultra. Scalable, high-performance edge computing built to accelerate enterprise AI across industries.
Compatible PCIe series
-
Qualcomm® Cloud AI 100 Ultra
-
NVIDIA RTX PRO 6000 Blackwell GPU
-
NVIDIA RTX 6000 Ada GPU
Specifications
| Model Number | AIP-FR68S-A1 | |
| CPU | Supports 12/13/14th Gen. Intel® Core i9/i7/i5 Processor up to 65W | |
| Grahics |
Intel® UHD Graphics 770 |
|
| Chipset | Intel® R680E Chipset | |
| System Memory | 4x DRR5 288-pin U-DIMM up to 192G | |
| Storage | 2x 2.5" SATAIII SSD (RAID 0/1) 1 x M.2 M-Key Slot, Support PCIe Gen4 x4 for NVME/SATA, Size 2280 1 x M.2 M-Key Slot, Support PCIe Gen4 x4 for NVME, Size 2280 |
|
| USB | 8 x USB 3.2 Gen2 x 1 (Type-A) 1 x USB 3.2 Gen2 x 2 (Type-C) |
|
| LAN | 2 x 2.5G RJ-45 Ethernet | |
| Serial port | 2 x RS-232/422/485 (COM1/2, DB9, Default RS-232) | |
| Display | 1 x HDMI 2.0 2 x DP++ 1.4 |
|
| Expansion Slots | 1x M.2 E-Key 2230 (USB2.0/PCIe) 1x M.2 B-Key 3052(USB2.0/USB3.2) GPU support：NVIDIA RTX PRO 6000 Blackwell, GeForce RTX 50 series, Qualcomm Cloud AI 100 Ultra |
|
| SIM Slot | 1x Nano SIM Slot | |
| Audio | 1 x Line-out, 1 x Mic-in | |
| Button | 1 x Power button | |
| Indicator | PCIe Slot-1&2 (POWER/FAN/TEMP) | |
| Power Consumption | Full loading : up to 1600 Watts | |
| Power Input /Connector | AC-in 110~220VAC | |
| Dimension (Wx Dx H) | 240 x 400 x 206 mm | |
| Mounting | Wall mount / Desk mount | |
| Net Weight | 12 kg | |
| Vibration | 1Grms ,IEC60068-2-64,Random,5 ~ 500 Hz ,1Hr / Axis (operation, w/o PCIe Card) | |
| Shock | 10G, IEC 60068-2-27, Half Sine, 11 ms Duration (operation, w/o PCIe Card) | |
| Temperature | Operating Temp.: 0°C ~ +50°C (32°F ~ 122°F)(w/o GPU) Operating Temp.: 0°C ~ +40°C (32°F ~ 122°F)(w/ GPU) Storage Temp.: -40°C ~ +85°C (- 40°F ~ 185°F) |
|
| Humidity | 95% @ 40°C Related Humidity, Non-condensing | |
| OS Support | Windows 11 IOT / Ubuntu 20.04/24.04 | |
| Security | TPM 2.0 | |
| Certification | CE / FCC Class A / UKCA / LVD / RoHS | |