Top
Author
Published
10 Nov 2026Form Number
LP2561PDF size
13 pages, 423 KBSubscribed to LP2561.
Thank you for your feedback.
Table of Contents
Abstract
The ThinkSystem NVIDIA ConnectX-9 9180 800Gb XDR/2x400GbE OSFP PCIe Gen6 x16 Adapter is a high-performance adapter that supports one port of 800Gb XDR InfiniBand or two ports of 400 Gb Ethernet, via a single-cage OSFP interface, providing best-in-class connectivity for AI computing.
This product guide provides essential presales information to understand the ConnectX-9 9180 adapter and its key features, specifications, and compatibility. This guide is intended for technical specialists, sales specialists, sales engineers, IT architects, and other IT professionals who want to learn more about the ConnectX-9 9180 adapter and consider its use IT solutions.
Introduction
The ThinkSystem NVIDIA ConnectX-9 9180 800Gb XDR/2x400GbE OSFP PCIe Gen6 x16 Adapter is optimized to supercharge hyperscale AI computing workloads. With support for both InfiniBand and Ethernet networking at up to 800 Gb/s, ConnectX-9 SuperNIC delivers high-speed, efficient network connectivity, significantly enhancing system performance for AI factories and cloud data center environments. The ConnectX-9 9180 adapter is available as a PCIe adapter or OCP adapter (PCIe adapter planned for January 2027).
The following figure shows the ConnectX-9 9180 adapter in PCIe form factor (shown with the full-adapter gold case and internal heatsink removed)
Figure 1. ThinkSystem NVIDIA ConnectX-9 9180 800Gb XDR/2x400GbE OSFP PCIe Gen6 x16 Adapter (full-adapter gold case and heatsink removed)
Did you know?
The ConnectX-9 9180 adapters are post-quantum cryptography (PQC) ready, providing security capabilities designed for the transition to cryptographic algorithms that can resist attacks from future quantum computers. The adapters support CNSA 2.0 post-quantum cryptography for secure boot and firmware update, helping protect the platform as security requirements evolve.
ConnectX-9 9180 provides up to 800 Gb/s connectivity using either InfiniBand or Ethernet, allowing the same adapter architecture to address different high-performance network infrastructures. With support for XDR InfiniBand and 800 GbE through a single OSFP connection, the ConnectX-9 9180 is well suited for the extremely high network bandwidth requirements of large-scale AI, HPC, and accelerated computing environments.
Part number information
The following table shows the ordering information.
The option part number includes the following:
- One NVIDIA adapter with full-height (3U) adapter bracket attached
- Low-profile (2U) adapter bracket
- Documentation
Supported transceivers and cables
The ConnectX-9 9180 adapter have an empty OSFP112 cage for connectivity.
The following table lists the supported transceivers.
The following table lists the supported optical cables.
The following table lists the supported direct-attach copper (DAC) cables.
Features
The NVIDIA ConnectX-9 9180 adapters have the following features:
-
Up to 800 Gb/s network connectivity: Provides a single OSFP network port supporting up to 800 Gb/s bandwidth. The 9180 adapters support both InfiniBand and Ethernet, enabling very high-bandwidth connectivity for AI, HPC, and cloud data center workloads.
-
InfiniBand and Ethernet support: Supports XDR, NDR, NDR200, HDR, and HDR100 InfiniBand connectivity, as well as Ethernet operating at 800, 400, 200, and 100 Gb/s.
-
PCIe 6.0 host interface: Uses a PCIe Gen6 x16 host interface, providing high-bandwidth connectivity between the adapter and host system. The interface operates at up to 64 GT/s per lane and is also compatible with PCIe Gen5.
-
Designed for accelerated AI networking: Includes hardware capabilities for RDMA and RoCEv2 acceleration, programmable RDMA transport, and advanced programmable congestion control, providing efficient network connectivity for demanding AI fabrics and accelerated computing environments.
-
NVIDIA GPUDirect acceleration: Supports NVIDIA GPUDirect RDMA and GPUDirect Storage, providing efficient data paths involving GPUs, network adapters, and storage. These capabilities are particularly applicable to GPU-intensive AI and HPC environments.
-
Advanced congestion control: Provides advanced programmable and telemetry-based congestion-control capabilities designed to maintain network efficiency for large-scale AI workloads.
-
In-Network Computing acceleration: Extends NVIDIA Scalable Hierarchical Aggregation and Reduction Protocol (SHARP) capabilities for In-Network Computing, helping accelerate supported collective operations for scientific computing, large-scale AI training, and inference workloads.
-
NVIDIA Spectrum-X integration: Supports NVIDIA Spectrum-X Ethernet technologies, including Spectrum-X multi-plane and the Spectrum-X net plug-in, for integration into large-scale accelerated computing environments.
-
Virtualization and cloud networking acceleration: Provides SR-IOV and NVIDIA Accelerated Switching and Packet Processing (ASAP²) capabilities for software-defined networking and virtual network functions. Acceleration capabilities include Open vSwitch, overlay networking, connection tracking, and network address translation (NAT).
-
TCP offload capabilities: Provides stateless TCP offloads to reduce host processing associated with networking functions and improve efficiency in cloud and data center networking environments.
-
Multi-Host connectivity: Supports NVIDIA Multi-Host technology for connectivity to as many as four hosts, enabling flexible connectivity configurations in supported systems.
-
Socket Direct capability: Supports NVIDIA Socket Direct technology, which can provide direct PCIe connectivity from separate CPUs in a dual-socket server to the network adapter, helping optimize network access in multi-socket server configurations.
-
Hardware-accelerated cryptography: Provides AES-GCM acceleration using 128-bit and 256-bit keys for data-in-motion security, including IPsec, TLS, and PSP, together with AES-XTS 256/512-bit key acceleration for data-at-rest protection.
-
Advanced platform security: Includes secure boot with a hardware root of trust, secure firmware update, flash encryption, and device attestation using SPDM 1.1 and SPDM 1.2. The 9180 adapters covered by this product guide are crypto enabled, secure boot enabled, and PQC ready.
-
Post-quantum cryptography (PQC) readiness: Supports CNSA 2.0 post-quantum cryptography capabilities for secure boot and firmware update. The 9180 adapters covered by this product guide are PQC ready.
-
Advanced timing and synchronization: Provides precision network timing and synchronization capabilities, including IEEE 1588 PTP, Precision Time Measurement (PTM), TCXO, PPS input/output, and SyncE.
Technical specifications
The ThinkSystem NVIDIA ConnectX-9 9180 800Gb XDR/2x400GbE OSFP PCIe Gen6 x16 Adapter have the following technical specifications:
Network Interface
- One OSFP112 cage for a transceiver or cable
- Max. bandwidth: 800 Gb/s
- 200/100/50 Gb/s of PAM4 and 25/10 Gb/s NRZ
Protocol support
- InfiniBand protocol support, based on InfiniBand Architecture Specification (IBTA) v1.7:
- 800 Gb/s XDR, (4 lanes × 200 Gb/s per lane)
- 400 Gb/s NDR (4 lanes × 100 Gb/s per lane)
- 200 Gb/s NDR200 (2 lanes × 100 Gb/s per lane)
- 100 Gb/s HDR100 (2 lanes × 50 Gb/s per lane)
- 50 Gb/s HDR (50 Gb/s per lane)
- Ethernet protocols supported are as follows:
- 800 Gb/s: 800GAUI-4 C2M, 800GAUI-8 C2M, 800GBASE-CR4, 800GBASE-CR8
- 400 Gb/s: 400GAUI-2 C2M, 400GAUI-4 C2M, 400GAUI-8 C2M, 400GBASE-CR2, 400GBASE-CR4, 400GBASE-CR8
- 200 Gb/s: 200GAUI-1 C2M, 200GAUI-2 C2M, 200GAUI-4, 200GAUI-4 C2M, 200GBASE-CR1, 200GBASE-CR2, 200GBASE-CR4
- 100 Gb/s: 100GAUI-1 C2M, 100GAUI-2 C2M, CAUI-4 C2M, 100GBASE-CR1, 100GBASE-CR2, 100GBASE-CR4
- 50 Gb/s: 50GAUI-1 C2M, 50GBASE-CR
- 25 Gb/s: 25GAUI C2M, 25GBASE-CR, 25GBASE-CR-S
Host Interface
- PCIe Gen6 (up to 48 lanes)
- NVIDIA Multi-Host (up to 4 hosts)
- PCIe switch downstream port containment (DPC)
- Message-signaled interrupts (MSI) and MSI-X
AI Networking
- Remote direct-memory access (RDMA) and RDMA over Converged Ethernet v2 (RoCEv2) accelerations
- NVIDIA Spectrum-X Ethernet:
- Spectrum-X multi-plane
- Spectrum-X net plug-in
- Programmable RDMA transport
- Advanced, programmable congestion control
- NVIDIA Collective Communications Library (NCCL)
- NVIDIA Inference Transfer Library (NIXL)
- NVIDIA GPUDirect RDMA
- GPUDirect Storage
- In-Network Computing
- Advanced timing and synchronization
- Message Passing Interface (MPI) accelerations
Cloud Networking
- Stateless Transmission Control Protocol (TCP) offloads
- Single-root IO virtualization (SR-IOV)
- NVIDIA Accelerated Switching and Packet Processing (ASAP2) Ethernet for software-defined networking (SDN) and virtual network functions (VNF):
- Open vSwitch (OVS) acceleration
- Overlay network accelerations
- Connection tracking (L4 firewall) and network address translation (NAT)
Security
- AES-GCM 128/256-bit key data-in-motion accelerations: Internet Protocol Security (IPsec), Transport Layer Security (TLS), Protocol for Secure Packet (PSP)
- AES-XTS 256/512-bit key data-at-rest acceleration
- Platform security:
- Secure boot with hardware root of trust
- Secure firmware update
- Flash encryption
- Device attestation (SPDM 1.1 and SPDM 1.2)
- Commercial National Security Algorithm Suite (CNSA) 2.0 post-quantum cryptography (PQC) Secure Boot and Update support
Management and Control
- Network Control Sideband Interface (NC-SI) (OCP adapter only)
- Management Component Transport Protocol (MCTP) over System Management Bus (SMBus) and PCIe Platform Level Data Model (PLDM) for:
- Monitor and Control DSP0248
- Firmware Update DSP0267
- Redfish Device Enablement DSP0218
- Field-Replaceable Unit (FRU) DSP0257
- Security Protocols and Data Models (SPDM) DSP0274
- Serial Peripheral Interface (SPI) to flash
Network Boot
- InfiniBand or Ethernet
- Preboot eXecution Environment (PXE) boot
- Unified Extensible Firmware Interface (UEFI)
NVIDIA Unified Fabric Manager
NVIDIA Unified Fabric Manager (UFM) is InfiniBand networking management software that combines enhanced, real-time network telemetry with fabric visibility and control to support scale-out InfiniBand data centers. The two UFM offerings are available from Lenovo:
- UFM Telemetry for Real-Time Monitoring
The UFM Telemetry platform provides network validation tools to monitor network performance and conditions, capturing and streaming rich real-time network telemetry information, application workload usage, and system configuration to an on-premises or cloud-based database for further analysis.
- UFM Enterprise for Fabric Visibility and Control
The UFM Enterprise platform combines the benefits of UFM Telemetry with enhanced network monitoring and management. It performs automated network discovery and provisioning, traffic monitoring, and congestion discovery. It also enables job schedule provisioning and integrates with industry-leading job schedulers and cloud and cluster managers, including Slurm and Platform Load Sharing Facility (LSF).
The following table lists the UFM subscription licenses available from Lenovo.
For more information on NVIDIA UFM, see the NVIDIA Unified Fabric Manager product page.
Support of adapters with generic firmware
One or more of the adapters described in this product guide uses standard vendor firmware (look for "Generic FW" or "Generic" in the adapter names). These adapters are supported in Lenovo servers and with Lenovo management tools.
Support in Lenovo XClarity management tools for adapters with generic firmware is per the following table.
Tip: Always use firmware that is obtained from Lenovo sources to ensure the firmware is fully tested by Lenovo and is supported. You should not use firmware that is obtained from the vendor web site, unless directed to do so by Lenovo support.
* Minimum supported UEFI firmware is as follows: SR630 V4, SR650 V4, SR650a V4: Release IHE116I; SR850 V4, SR860 V4: Release RVE106I.
Regulatory approvals
The adapter has the following regulatory approvals:
- Safety: CB / cTUVus / CE
- EMC: CE / FCC / VCCI / ICES / RCM / KC
- RoHS: RoHS Compliant
Operating environment
The adapter has the following operating characteristics:
- Maximum power available through OSFP port: 17W
- Temperature
- Operational: 0°C to 55°C
- Non-operational: -40°C to 70°C
- Humidity: 90% relative humidity
Warranty
One year limited warranty. When installed in a Lenovo server, the adapter assumes the server’s base warranty and any warranty upgrades.
Related publications
For more information, refer to these documents:
- ThinkSystem Ethernet and InfiniBand Adapter Reference
https://lenovopress.lenovo.com/lp1594-thinksystem-ethernet-infiniband-adapter-reference - Lenovo ServerProven compatibility information
http://serverproven.lenovo.com - NVIDIA ConnectX InfiniBand product page:
https://www.nvidia.com/en-us/networking/infiniband-adapters/ - ConnectX-9 SuperNIC user manual:
https://networking-docs.nvidia.com/connectx9hw/ - ConnectX-9 SuperNIC for OCP 3.0 user manual:
https://networking-docs.nvidia.com/connectx9ocphw/
Trademarks
Lenovo and the Lenovo logo are trademarks or registered trademarks of Lenovo in the United States, other countries, or both. A current list of Lenovo trademarks is available on the Web at https://www.lenovo.com/us/en/legal/copytrade/.
The following terms are trademarks of Lenovo in the United States, other countries, or both:
Lenovo®
ServerProven®
ThinkSystem®
XClarity®
The following terms are trademarks of other companies:
AMD® is a trademark of Advanced Micro Devices, Inc.
Intel®, the Intel logo is a trademark of Intel Corporation or its subsidiaries.
NVIDIA®, NVIDIA Socket Direct, NVIDIA Spectrum®, and NVIDIA Spectrum-X are trademarks of NVIDIA Corporation.
Other company, product, or service names may be trademarks or service marks of others.
Configure and Buy
Please select a locale
Full Change History
Course Detail
Employees Only Content
The content in this document with a is only visible to employees who are logged in. Logon using your Lenovo ITcode and password via Lenovo single-signon (SSO).
The author of the document has determined that this content is classified as Lenovo Internal and should not be normally be made available to people who are not employees or contractors. This includes partners, customers, and competitors. The reasons may vary and you should reach out to the authors of the document for clarification, if needed. Be cautious about sharing this content with others as it may contain sensitive information.
Any visitor to the Lenovo Press web site who is not logged on will not be able to see this employee-only content. This content is excluded from search engine indexes and will not appear in any search results.
For all users, including logged-in employees, this employee-only content does not appear in the PDF version of this document.
This functionality is cookie based. The web site will normally remember your login state between browser sessions, however, if you clear cookies at the end of a session or work in an Incognito/Private browser window, then you will need to log in each time.
If you have any questions about this feature of the Lenovo Press web, please email David Watts at dwatts@lenovo.com.
