July 8, 2025
news

Premio Inc. has released LLM-1U-RPL, the first entry for the new LLM series line edge server. The compact, short concentration (483(w) x 480(d) x 44(h)mm) 1U edge server is designed to deliver real-time, generated AI (genai) and leading language model (LLM) workloads straight to on-premises data centers.
“The LLM-1U-RPL is dedicated to on-premises data centers and provides high-performance, low-latency AI inference for large-scale language model (LLM) workloads.
- 13th Generation Intel Core Processor (Maximum I9, 65W TDP)
- Support up to NVIDIA RTX 5000 ADA GPU for accelerated computing
- PCIEGEN 4 expansion for GPU AI accelerators or high-throughput network cards
- Flexible and fast storage options for M.2 NVME and dual hot swappable 2.5-inch SATA bays
- On-premises Optimized I/O Connection Edge AI: 3X 2.5GBE LAN port, 6x USB 3.2 Gen2 port, and COM port
- 600W (1+1) Redundant Power Supply
- Hot-swappable redundant smart fan
- Enhanced cybersecurity and physical security
- World-class certification (UL, FCC, CE)
LLM-1U-RPL supports up to 64GB of dual channel DDR4 3200MT/S SODIMM memory to streamline multimodal data streams without performance holdup. Includes high-speed NVME via M.2 M key slots and dual hot-swappable 2.5-inch SATA bays with front access.
The server provides PCIE GEN 4 expansion slots for high-throughput network interface cards (NICs) or dedicated AI GPU accelerators, enabling high-performance inference for private, on-site LLM deployments, including digital twins and generated AI inference. Ideal for automation and robotics, smart infrastructure and security manufacturing.
“Designed for the demand for edge deployment, this new Edge server integrates a 13th GenIntel® Core™ processor with a performance hybrid architecture, as well as a dedicated NVIDIA GPU for accelerated computing, enabling real-time intelligence, reducing potential and giving you more powerful control over your data.
For more information, see premioinc.com/collections/llm-1u-rpl-series-edge-ai-rackmount-server.
