NVIDIA Quantum-2 QM9700 NDR 400G Managed InfiniBand Switch
Managed NDR 400G core switch with embedded subnet manager for GPU clusters.
We help you choose, configure, and deliver the right system — no obligation.
Configuration at a Glance
Tailored per engagement. Full technical overview below.
Overview
The Quantum-2 QM9700 is a 1U internally managed NDR InfiniBand switch delivering 64 ports of 400Gb/s across 32 OSFP cages, with an on-board subnet manager for turnkey fabric bring-up. Nexus Compute specifies, configures, and validates each QM9700 with the correct firmware and airflow option, delivered warranty-backed through NVIDIA-authorized channels.
Who This Solution Is For
Business Benefits
Turnkey fabric management
The embedded subnet manager brings up clusters of up to roughly 2,000 nodes without a separate management appliance.
Non-blocking 51.2 Tb/s
Full bisectional bandwidth keeps GPUs saturated during collective-heavy training runs.
Validated and warranty-backed
Each unit is configured, tested, and sourced through authorized channels with full manufacturer warranty.
Typical Business Use Cases
Core/spine switching for AI training clusters
Self-managed HPC InfiniBand fabrics
GPU-to-GPU RDMA interconnect
Lossless storage and compute backbone
Industry Applications
Technical Overview
Built on the NVIDIA Quantum-2 ASIC, the QM9700 provides 64 NDR 400Gb/s ports on 32 physical OSFP connectors, splittable to 128 NDR200 ports, with 66.5 billion packets-per-second forwarding capacity. In-network SHARPv3 acceleration offloads collective operations directly into the fabric to cut MPI and all-reduce latency.
| Switch ASIC | NVIDIA Quantum-2 |
| Ports | 64x NDR 400Gb/s (32x OSFP), to 128x NDR200 |
| Aggregate Throughput | 51.2 Tb/s bidirectional |
| Forwarding Capacity | 66.5 Bpps |
| Management | Embedded subnet manager (up to ~2,000 nodes), MLNX-OS |
| In-Network Computing | SHARPv3 collective acceleration |
| Form Factor | 1U rack-mount |
| Power / Airflow | Dual hot-swap PSUs, P2C or C2P airflow option |
| Warranty | Manufacturer-backed via authorized channels |
Specifications are indicative and configured to each engagement. Request a quote for a configuration tailored to your requirements.
Warranty, Support & Fulfillment
Every system ships from an authorized channel, configured and tested, with the documentation enterprise buyers need — backed by warranty and a dedicated account team.
Enterprise Warranty
Full manufacturer warranty with optional on-site, next-business-day support and extended coverage.
Authorized Channel
Sourced through Tier-1 distribution and OEM partners — never grey market. Asset & warranty records included.
Lead Time & Deployment
48-hour quotes, then configured, burn-in tested, and delivered on a committed schedule.
Nationwide Fulfillment
Coordinated logistics, rack-and-stack, and delivery wherever your infrastructure lives.
Frequently Asked Questions
Do I need a separate subnet manager with the QM9700?
No. The QM9700 is internally managed with an embedded subnet manager that brings up fabrics of up to roughly 2,000 nodes, so smaller clusters need no external management appliance.
How many GPU nodes can a single QM9700 support?
With 64 NDR 400G ports, a single switch can serve as a non-blocking leaf or spine for dozens of dual-rail nodes; we size leaf/spine counts to your exact node and rail topology.
QM9700 or QM9790 for my fabric?
Choose the managed QM9700 for self-contained clusters; choose the unmanaged QM9790 when you standardize on external NVIDIA UFM management across a larger fabric.
Hardware Assistance
Configure the NVIDIA Quantum-2 QM9700 NDR 400G Managed InfiniBand Switch with Nexus Compute
Tell us your requirements and a hardware specialist will help you specify, configure, and quote the right system — typically within two business days. No obligation.