Short-depth LLM-2U-AM5 and LLM-3U-AM5 servers deliver AMD processing, GPU acceleration, and PCIe Gen 5 expansion for private AI and edge computing.
LOS ANGELES, CA, UNITED STATES, November 12, 2026 /EINPresswire.com/ — Short-Depth 19-Inch Rackmount Servers With AMD AM5 Platform for Physical AI and On-Premises AI
As organizations move AI processing closer to where data is generated, demand is growing for compact, high-performance computing infrastructure beyond traditional data centers. The LLM-2U-AM5 and LLM-3U-AM5 address these needs with short-depth 2U and 3U rackmount designs measuring just 361 mm (14.2 inches) in chassis depth, bringing server-class AMD processing into space-constrained environments. Designed for standard 19-inch racks, both platforms enable space-efficient integration into on-premises infrastructure while supporting physical AI, private AI, localized LLM inference, robotics, and distributed edge computing workloads.
Built on the AMD AM5 platform, both servers support AMD EPYC™ 4005 and Ryzen™ 9000 Series processors, scaling up to 16 cores and 32 threads. PCIe Gen 5 expansion supports compatible AI accelerators, while dual Broadcom 10GbE SFP+ networking helps reduce data bottlenecks between cameras, storage and edge compute nodes. Together, these capabilities enable IT and OT teams to process data locally, support demanding AI workloads and reduce reliance on centralized cloud infrastructure.
With up to 192GB DDR5 memory, flexible M.2 and SATA storage, and dual displays, the platforms provide versatile integration options for industrial and AI applications. Premio’s remote management capabilities support system monitoring, troubleshooting and recovery across distributed deployments, helping IT and OT teams reduce on-site maintenance and simplify infrastructure management.
Key Features
• Supports AMD Ryzen™ Embedded 9000 Series & EPYC™ 4005 Processors (Max. 170W TDP)
• AMD B650 Chipset
• 19-inch Short-Depth Rackmount Design, Standard 2U and 3U Form Factors
• 4x DDR5 UDIMM, up to 5200 MT/s, Max. 192GB (Non-ECC)
• 2x 10GbE SFP+, 2x GbE LAN
• Dual Display: 1x DP, 1x HDMI
• 2x PCIe x16 Gen 5, 1x PCIe x1
• 2x M.2 M-Key, 1x M.2 E-Key
• SATA SSD Drive Bays (Support RAID 0/1/5/10): 4x 2.5″ (LLM-2U-AM5) / 4x 2.5″ or 1x 3.5″ (LLM-3U-AM5)
• 6x USB 3.2 Gen 2 (Rear), 2x USB 3.2 (Front)
• Optional Out-of-Band (OOB) Remote Management
• 850W Single Power Supply
• Supports TPM 2.0
• UL 62368 Ed. 4, CE, FCC
Scaling Physical AI Infrastructure with the LLM Series Short-Depth 19-Inch Rackmount Servers
With the introduction of the LLM-2U-AM5 and LLM-3U-AM5, Premio’s LLM Series now spans 1U, 2U, and 3U short-depth 19-inch rackmount servers, offering a scalable portfolio for distributed Edge AI infrastructure. From compact 1U systems for space-constrained deployments to 2U and 3U platforms offering greater computing and expansion flexibility, the LLM Series enables IT and OT engineers to match processing performance, PCIe expansion, and infrastructure requirements to their specific robotics AI workloads.
Powered by Intel® and AMD processor platforms across the family, the LLM Series supports a broad range of on-premises AI applications. The new LLM-2U-AM5 and LLM-3U-AM5 bring server-class AMD multicore processing, PCIe expansion, and high-speed networking to support localized AI inference, private AI, robotics, machine vision, warehouse automation, and Physical AI workloads.
Designed for standard 19-inch racks, the LLM Series integrates into existing IT/OT infrastructure, supporting centralized power distribution, UPS integration, organized airflow management, secure equipment enclosures, and simplified maintenance. This standardized rackmount architecture enables organizations to scale distributed Edge AI infrastructure while maintaining greater control over computing resources, data processing, and long-term system operations.
Dustin Seetoo
Premio Inc.
+1 626-839-3100
Visit us on social media:
LinkedIn
YouTube
Premio’s LLM Series: Short-Depth Edge AI Rackmount Servers
Legal Disclaimer:
EIN Presswire provides this news content “as is” without warranty of any kind. We do not accept any responsibility or liability
for the accuracy, content, images, videos, licenses, completeness, legality, or reliability of the information contained in this
article. If you have any complaints or copyright issues related to this article, kindly contact the author above.
Media gallery
