1 1

Cisco

Cisco UCSAI-880A-M8-CHS | UCS C880A M8 10RU 8-GPU AI Server Chassis, HGX B300

UCSAI-880A-M8-CHS

Stock Status: Enquire

Request Quote
Sale Sold out
Shipping calculated at checkout.
20-min response

Need a quote or bulk pricing? Answered within 20 minutes, Sydney business hours.

Description

The UCS C880A M8 is a high-density, air-cooled 10RU rack server platform built on the NVIDIA HGX reference architecture. It is aimed at AI and HPC deployments, covering large-scale model training, fine-tuning, real-time LLM inference, RAG, GPU-accelerated ETL and big-data analytics in AI factories and Cisco AI PODs. This listing is the chassis-level (L10) part for the 8-GPU HGX B300 platform within the UCS C-Series family.

The platform is designed around eight NVIDIA HGX B300 NVL8 GPUs with 2.3 TB of total GPU memory and two 6th Gen Intel Xeon processors in configured servers. Memory scales to 32 DDR5-6400 DIMM slots and up to 4 TB. Eight board-integrated NVIDIA ConnectX-8 NICs carry east-west traffic, while four PCIe Gen5 x16 FHHL slots and an OCP 3.0 slot handle north-south networking. Local storage supports up to 8 E1.S PCIe Gen5 NVMe SSDs plus two M.2 SATA boot drives.

Power is delivered by twelve 3200W 50V Titanium-grade hot-swap supplies in N+N redundancy, with 20 hot-swap fans in N+1. System power consumption reaches up to 14.3 kW, so rack power and cooling planning is essential. Optional parts for the family include DDR5 RDIMMs, E1.S NVMe drives, ConnectX-7 and BlueField-3 network adapters, a security bezel and a ball-bearing rail kit.

Enquire with XS Network Tech for new or used pricing and availability.

Features

- NVIDIA HGX platform with 8x B300 SXM GPUs and high-speed GPU interconnect
- Two 6th Gen Intel Xeon CPUs to feed GPU training and inference pipelines
- Embedded 800G ConnectX-8 east-west fabric for low-latency scale-out AI training
- Local E1.S NVMe storage for caching and fast checkpointing
- Hot-swappable N+N power supplies and N+1 fans for availability
- Part of Cisco validated AI solutions and Cisco AI PODs
- Supports NVIDIA AI Enterprise and NVIDIA NIM

Warranty

All products sold by XS Network Tech include a 12-month warranty on both new and used items. Our in-house technical team thoroughly tests used hardware prior to sale to ensure enterprise-grade reliability.

All technical data should be verified on the manufacturer data sheets.

View full details

specs-tabs

Technical Specifications

FAQs

Technical Specifications

- Form factor: 10RU 19-inch rack server (NVIDIA HGX reference architecture)
- GPU: 8x NVIDIA HGX B300 NVL8 (2.3 TB total GPU memory) in configured servers
- Processors: 2x 6th Gen Intel Xeon (6776P or 6767P) in configured servers
- Memory: 32 DDR5-6400 DIMM slots, up to 4 TB (32x 128GB)
- Boot drive: 2x 960GB M.2 SATA SSD with RAID controller
- Internal storage: up to 8x PCIe Gen5 x4 E1.S NVMe SSD
- East-west networking: 8x GPU-board integrated NVIDIA ConnectX-8
- North-south networking: 2x ConnectX-7 2x200G, 2x B3220 2x200G or 2x B3240 2x400G, 1x OCP TFF Gen5 x8
- Expansion: 4x PCIe Gen5 x16 FHHL slots and 1x OCP 3.0 TFF NIC (Intel X710-T2L, 2x 10GbE RJ-45)
- Power supply: 12x 3200W 50V MCRPS hot-swap PSUs (N+N)
- System power consumption: up to 14.3 kW
- Input rating: 200-277 Vac, 50-60 Hz, 16A per PSU
- Cooling: 20x hot-swappable fans (N+1)
- Management: Cisco BMC via DC-SCM (RJ45, IPMI 2.0, KVM with dedicated LAN)
- Front I/O: 2x USB 2.0, ID button, power button
- Rear I/O: 1x USB 3.0, 1x mDP, 1x USB 2.0 UART, 1x RJ45 management
- BIOS: AMI BIOS
- Dimensions (H x W x D): 441 x 447 x 800 mm
- Weight: 164 kg
- Operating temperature: 5 to 35 C, 5% to 85% non-condensing humidity

FAQs

Q: Does the chassis need a licence?
A: No licence is needed for the chassis hardware itself. NVIDIA AI Enterprise is an optional subscription sold per GPU in 1, 3 or 5-year terms.

Q: What power does this server need?
A: It uses twelve 3200W 50V Titanium-grade PSUs in N+N redundancy, rated 200-277 Vac at 16A per PSU. System consumption is up to 14.3 kW, and a power cord such as CAB-AC-16A-AUS (Australia, 16A, C19) must be selected.

Q: What options are available for this platform?
A: The family offers 64, 96 and 128GB DDR5-6400 RDIMMs, 3.84, 7.68 and 15.36TB E1.S NVMe drives, NVIDIA ConnectX-7 and BlueField-3 B3220 and B3240 adapters, an Intel X710-T2L OCP NIC, a security bezel (UCSAI-880A-BZL) and a ball-bearing rail kit (UCSAI-RAIL).

Q: How does this SKU differ from the configured UCS C880A M8 servers?
A: UCSAI-880A-M8-CHS is the chassis-level L10 part for the 8-GPU HGX B300 platform, whereas the configured UCSC-880A-M8-B30x SKUs are fully built servers.

Q: What networking does the platform provide?
A: Eight board-integrated NVIDIA ConnectX-8 NICs handle east-west traffic. North-south connectivity uses PCIe Gen5 x16 slots for ConnectX-7 or BlueField-3 adapters, plus an OCP 3.0 slot.

Related Products

View all Enterprise Servers: Rack, Tower & Blade

Recently Viewed

  • Request a Quote

    Looking for competitive pricing? Submit a request, and our team will provide a tailored quote that fits your needs.

  • Contact Us Directly

    Have a question or need immediate assistance? Call us for expert advice and real-time support.

    Call us Now  

  • Contact Us Directly

    Have a question or need immediate assistance? Call us for expert advice and real-time support.

    Contact us