NVIDIA Quantum-2 QM9700
The NVIDIA Quantum-2 QM9700 is a 1U managed NDR InfiniBand switch delivering 64 ports of 400 Gb/s NDR InfiniBand OSFP. Delivering 51.2 Tbps of non-blocking throughput with nanosecond latency, the QM9700 features an embedded Subnet Manager, SHARPv3 in-network computing, and adaptive routing. Each 400G OSFP port can be split into two 200G NDR200 links, supporting up to 128 ports in 1U. The QM9700 represents the apex of modern InfiniBand clustering for generative AI and LLM training supercomputers.
Specifications
| Brand | NVIDIA |
|---|---|
| Family | Quantum |
| Form factor | 1U fixed switch |
| Released | Nov 2021 |
| Ports | 64x 400 Gb/s NDR InfiniBand OSFP ports (or up to 128x 200 Gb/s NDR200 ports) |
| Switching capacity | 51.2 Tbps aggregate non-blocking switching throughput |
| Silicon | NVIDIA Quantum-2 7nm InfiniBand processing ASIC |
| In network computing | NVIDIA SHARPv3 with FP16/FP32/FP64 in-network tensor acceleration |
| Management | Internally managed with embedded Subnet Manager supporting up to 2,000 nodes |
Generations
- Previous generation: NVIDIA Quantum QM8700
Compatible parts
Cables
MCP4Y00-N003 NVIDIA NDR 400Gb/s OSFP Direct Attach Copper Cable 3m
Transceivers
MMA4Z00-NS NVIDIA NDR 400Gb/s OSFP Optical Transceiver
Works together
How the NVIDIA Quantum-2 QM9700 connects to servers, storage and the network: 1 way, with the parts and products on each path.
GPU cluster fabric
InfiniBand adapters and switches must match the generation: HDR is 200Gb, NDR is 400Gb. VPI adapters such as ConnectX-6 and ConnectX-7 can run either InfiniBand or Ethernet.
Servers
Dell PowerEdge XE8545 4U rack
Lenovo ThinkSystem ST650 V2 tower
Lenovo ThinkSystem SR670 V2 3U rack
HPE Apollo 6500 Gen10 Plus 4U rack
HPE ProLiant DL385 Gen11 2U rack
Lenovo ThinkSystem SR860 V2 4U rack
GPUs
NVIDIA Data Center GPU A10 PCIe single-slot full-height full-length
NVIDIA Data Center GPU A100 PCIe 80GB PCIe dual-slot full-height full-length
NVIDIA Data Center GPU A30 PCIe dual-slot full-height full-length
NVIDIA Data Center GPU H100 SXM5 80GB SXM5 module
Adapters for the server
MCX556A-ECAT NVIDIA ConnectX-5 VPI Dual-Port 100Gb/s EDR InfiniBand and 100GbE QSFP28 PCIe 3.0 x16 Adapter
MCX75310AAS-NEAT NVIDIA ConnectX-7 Single-Port 400Gb/s NDR InfiniBand and 400GbE OSFP PCIe 5.0 x16 Adapter
MCX755106AS-HEAT NVIDIA ConnectX-7 Dual-Port 200Gb/s NDR200 InfiniBand and 200GbE QSFP112 PCIe 5.0 x16 Adapter
540-BCVE Dell NVIDIA Mellanox ConnectX-6 VPI Single Port 200Gb/s HDR InfiniBand / 200GbE PCIe Adapter
Plan and size it
- Draw this setup in the topology planner mrplanb.com
- Tokens per kWh calculator for GPU racks rack-density.com
- Add up rack power and space rack-density.com
- Power, cooling and racks for dense GPU servers mrplanb.com
- Network monitoring tools compared mrplanb.com
Check compatibility with NVIDIA
Before you buy, confirm the exact configuration against the vendor's own lists: firmware levels, operating systems, adapters and optics that NVIDIA has tested with the NVIDIA Quantum-2 QM9700.
- Compatibility NVIDIA Qualified System Catalog (certified GPU servers)
- Specifications NVIDIA Networking documentation (adapters, switches, cables)
- Support NVIDIA Enterprise Support
FAQ
What optical form factor is used on the QM9700?
It utilizes 400 Gb/s OSFP (Octal Small Form-Factor Pluggable) transceivers and cables.
How does SHARPv3 accelerate LLM training?
SHARPv3 processes AllReduce tensor operations directly on the switch fabric, significantly speeding up distributed deep learning training.
Buying or selling Quantum-2 QM9700?
New, refurbished or used. Tell us the quantity and condition and we connect you with a vetted reseller in your region, usually within one working day.
Sources
- NVIDIA Quantum-2 QM9700 Product Overview, checked 2026-10-27
- NVIDIA Quantum-2 QM9700 Datasheet, checked 2026-10-27