 
|
NVIDIA C8180 ConnectX-8 SuperNIC 800Gb/s OSFP Adapter | PCIe Gen6 x16
Product Details:
| Brand Name: | Mellanox |
| Model Number: | C8180(900-9X81E-00EX-ST0) |
| Document: | connectx-datasheet-connectx...1).pdf |
Payment & Shipping Terms:
| Minimum Order Quantity: | 1pcs |
|---|---|
| Price: | Negotiate |
| Packaging Details: | outer box |
| Delivery Time: | Based on inventory |
| Payment Terms: | T/T |
| Supply Ability: | Supply by project/batch |
|
Detail Information |
|||
| Model NO.: | C8180(900-9X81E-00EX-ST0) | Transmission Rate: | 400gbe |
|---|---|---|---|
| Ports: | Dual-Port | Function: | Lacp, Stackable, Vlan Support |
| Technology: | Infiniband | Host Interface: | Gen6 X16 |
| Interface Type: | Qsfp112 | Trademark: | Mellanox |
| Highlight: | NVIDIA ConnectX-8 SuperNIC 800Gb/s adapter,PCIe Gen6 x16 network card,OSFP adapter with 800Gb/s speed |
||
Product Description
NVIDIA C8180 ConnectX-8 SuperNIC 800Gb/s OSFP Adapter
PCIe Gen6 x16 · InfiniBand / Ethernet · GPUDirect® · SHARP In-Network Computing
Max throughput: 800Gb/s (1*800G or 2*400G)
Host interface: PCIe Gen6 (up to 48 lanes)
Form factor: PCIe HHHL 1P (OSFP)
Security: Secure boot, flash encryption, IPsec, MACsec
In-Network Computing: SHARP, collective offloads, GPUDirect
Timing: IEEE 1588v2 Class C, SyncE, PTM
Product Overview
The NVIDIA ConnectX-8 C8180 SuperNIC (ordering part number: C8180-900-9X81E-00EX-ST0) is the industry's highest-performance 800Gb/s networking adapter, purpose-built for massive-scale AI and hyperscale cloud data centers. Supporting both InfiniBand and Ethernet protocols, it delivers extremely fast, efficient network connectivity that significantly enhances system performance for AI factories, trillion-parameter GPU computing, and agentic AI applications.
With PCIe Gen6 host interface, advanced routing, and telemetry-based congestion control, ConnectX-8 SuperNIC achieves peak AI workload efficiency. It seamlessly integrates with next-generation NVIDIA networking platforms, providing up to 800 Gb/s of end-to-end connectivity while improving power efficiency for sustainable AI data centers operating hundreds of thousands of GPUs.
Key Features
- 800Gb/s total bandwidth (InfiniBand / Ethernet)
- PCIe Gen6 host interface (up to 48 lanes)
- 1-port OSFP (800/400/200/100Gb/s)
- NVIDIA GPUDirect® RDMA & Storage
- SHARP in-network computing acceleration
- Advanced congestion control & telemetry
- Secure boot with hardware root-of-trust
- In-line IPsec, MACsec, PSP acceleration
- IEEE 1588v2 Class C (12 ns) & SyncE
- ASAP2 SDN / VNF acceleration
- RoCEv2 & RDMA acceleration
- NVIDIA Multi-Host™ (up to 4 hosts)
Technology & Acceleration Engines
NVIDIA In-Network Computing
ConnectX-8 SuperNIC extends NVIDIA SHARP (Scalable Hierarchical Aggregation and Reduction Protocol) to boost in-network computing in HPC and AI environments. Collective operations, rendezvous protocol, and burst-buffer offloads are hardware-accelerated, significantly improving efficiency for both training and inferencing at scale.
GPUDirect® RDMA
Enables direct data movement between GPU memory and the network, bypassing the CPU for ultra-low latency and high throughput in AI workloads.
ASAP2 (Accelerated Switch and Packet Processing)
Delivers line-rate SDN and VNF acceleration with overlay networking (VXLAN, GENEVE, NVGRE), connection tracking (L4 firewall), NAT, hierarchical QoS, and flow mirroring – all without CPU overhead.
Advanced Congestion Control
Programmable congestion control combined with telemetry-based routing ensures optimal network performance even under the most demanding AI traffic patterns.
Typical Deployments
- Massive-Scale AI Training – Trillion-parameter GPU clusters with GPUDirect and SHARP acceleration.
- AI Factories – High-throughput, low-latency fabrics for large-scale generative AI and foundation models.
- Hyperscale Cloud Data Centers – 800Gb/s connectivity for cloud-native AI workloads and multi-tenant environments.
- High-Performance Computing (HPC) – MPI and collective communication offloads for scientific simulations.
- Agentic AI Applications – Real-time inference and distributed reasoning across thousands of GPUs.
Compatibility
InfiniBand: 800/400/200/100Gb/s
Ethernet: 400/200/100/50/25Gb/s RoCEv2
Supports dual-protocol operation (InfiniBand and Ethernet) with auto-negotiation. Compatible with next-gen NVIDIA networking platforms, Quantum-3 InfiniBand switches, and industry-standard 800GbE / 400GbE switches.
Host Interface – PCIe Gen6 (up to 48 lanes) with bifurcation support, NVIDIA Multi-Host (up to 4 hosts), SR-IOV, ATS, PASID, and DPC.
Operating Systems – In-box drivers for Linux (RHEL, Ubuntu), Windows, VMware ESXi (SR-IOV), and Kubernetes.
Specifications
| Parameter | Details |
|---|---|
| Product model | C8180-900-9X81E-00EX-ST0 |
| Product family | NVIDIA ConnectX-8 SuperNIC |
| Network protocols | InfiniBand (IBTA v1.7) · Ethernet |
| Maximum total bandwidth | 800 Gb/s |
| InfiniBand speeds | 800/400/200/100 Gb/s |
| Ethernet speeds | 400/200/100/50/25 Gb/s |
| Port configuration | 1-port OSFP (supports split up to 8 ports) |
| Host interface | PCIe Gen6 (up to 48 lanes) · Multi-Host (4 hosts) |
| Form factor | PCIe HHHL 1P |
| I/O channels | 16 million |
| MTU | 256 to 4,096 bytes |
| Security offloads | IPsec, MACsec, PSP · Secure boot · Flash encryption · Device attestation (SPDM 1.1) |
| Timing & synchronization | IEEE 1588v2 Class C (12 ns) · SyncE · PTM · PPS in/out |
| SDN acceleration | ASAP2: VXLAN, GENEVE, NVGRE, L4 firewall, NAT, QoS, flow mirroring |
| In-Network Computing | SHARP · collective offloads · rendezvous offload · burst buffer · enhanced atomics |
| Virtualization | SR-IOV · VirtIO · Multi-Host (4 hosts) |
| Management | NC-SI · MCTP · PLDM (monitor, firmware, FRU, Redfish) · SPDM · SPI · JTAG |
| Remote boot | InfiniBand / Ethernet · PXE · iSCSI · UEFI |
Selection Guide
The ConnectX-8 SuperNIC family offers multiple form factors and port configurations to suit diverse deployment needs.
| Ordering Part Number | Ports | Form Factor | Host Interface | Description |
|---|---|---|---|---|
| C8180-900-9X81E-00EX-ST0 | 1x OSFP | PCIe HHHL 1P | PCIe Gen6 | 800Gb/s single-port, InfiniBand/Ethernet SuperNIC |
| C8180-900-9X81E-00EX-ST1 | 2x QSFP112 | PCIe HHHL 2P | PCIe Gen6 | 800Gb/s dual-port (2*400G), QSFP112 |
| C8180-900-9X81E-00EX-OCP | 1x OSFP | OCP 3.0 TSFF | PCIe Gen6 | OCP 3.0 form factor for open rack |
| C8180-900-9X81E-00EX-OCP | 2x QSFP112 | Dual ConnectX-8 Mezzanine | PCIe Gen6 | Mezzanine for dense GPU platforms |
Contact Starsurge for volume pricing, custom SKUs, and lead time.
Advantages
- Unmatched bandwidth – 800Gb/s for massive-scale AI and data-intensive workloads.
- Lowest latency – GPUDirect and in-network computing minimize data movement overhead.
- Future-ready – PCIe Gen6 and 800G readiness for next-gen AI infrastructure.
- Enhanced security – hardware root-of-trust, secure boot, and in-line encryption at line rate.
- Power efficient – optimized for sustainable AI data centers operating hundreds of thousands of GPUs.
- Unified fabric – single adapter for InfiniBand and Ethernet, reducing complexity.
Service & Support
Hong Kong Starsurge Group offers comprehensive support for NVIDIA ConnectX-8 SuperNICs:
- Technical consultation for AI fabric design and performance optimization.
- Global logistics with fast delivery from regional hubs.
- Firmware update and secure boot configuration assistance.
- 3-year warranty with advanced replacement.
FAQ
What is the maximum throughput of the ConnectX-8 SuperNIC?
It supports up to 800 Gb/s total bandwidth, configurable as 1*800G or 2*400G depending on the port configuration.
Does it support both InfiniBand and Ethernet?
Yes, it supports dual-protocol operation with auto-negotiation between InfiniBand and Ethernet.
What is NVIDIA SHARP and how does it benefit AI workloads?
SHARP (Scalable Hierarchical Aggregation and Reduction Protocol) offloads collective communication operations to the network, significantly accelerating distributed training and inference.
Is this adapter compatible with PCIe Gen5 systems?
Yes, it is backward compatible with PCIe Gen5 and Gen4, though maximum bandwidth will be limited by the host interface.
What security features are included?
Secure boot with hardware root-of-trust, secure firmware update, flash encryption, device attestation (SPDM 1.1), and in-line IPsec/MACsec/PSP acceleration.
Handling & Precautions
- Use ESD-safe practices when handling the adapter – avoid touching the gold fingers.
- Ensure adequate airflow in the server chassis for proper cooling of the high-performance adapter.
- Do not force the adapter into the PCIe slot; align carefully.
- Firmware updates should be performed per NVIDIA guidelines to avoid disruption.
- For security features (secure boot, encryption), follow the platform security enablement guide.
About Hong Kong Starsurge Group
Hong Kong Starsurge Group Co., Limited is a technology-driven provider of network hardware, IT services, and system integration solutions. Founded in 2008, we serve customers worldwide with products including network switches, NICs, wireless access points, controllers, cables, and related equipment. With an experienced sales and technical team, Starsurge supports industries such as government, healthcare, manufacturing, education, finance, and enterprise. We also offer IoT solutions, network management systems, custom software development, multilingual support, and global delivery. Our customer-first approach ensures reliable quality, responsive service, and tailored solutions that help clients build efficient, scalable, and dependable network infrastructure.
Key Facts
- 800Gb/s single-port OSFP
- PCIe Gen6 (up to 48 lanes)
- InfiniBand & Ethernet dual-protocol
- GPUDirect RDMA + Storage
- SHARP in-network computing
- Secure boot + flash encryption
- IEEE 1588v2 Class C (12 ns)
- ASAP2 SDN acceleration
- RoCEv2 & RDMA acceleration
- NVIDIA Multi-Host (4 hosts)
Compatibility Matrix
| System | Port Type | Speed | Protocol |
|---|---|---|---|
| NVIDIA Quantum-3 (next-gen) | OSFP | 800Gb/s | InfiniBand |
| NVIDIA Spectrum-4 / Spectrum-5 | OSFP | 800GbE / 400GbE | Ethernet |
| DGX / HGX platforms | OSFP | 800Gb/s | InfiniBand / RoCE |
| Third-party 800G switches | OSFP | 800GbE | Ethernet (RoCE) |
Buyer Checklist
- Confirm PCIe slot: Gen6 x16 (backward compatible with Gen5/Gen4)
- Verify form factor: HHHL 1P fits your chassis
- Check OS/driver support (Linux, Windows, VMware, Kubernetes)
- Determine protocol: InfiniBand or Ethernet (dual-protocol capable)
- Evaluate security requirements (secure boot, IPsec, MACsec)
Related Products
- NVIDIA ConnectX-8 SuperNIC OCP 3.0 (C8180-900-9X81E-00EX-OCP)
- NVIDIA ConnectX-8 Dual-Port Mezzanine
- NVIDIA Quantum-3 InfiniBand switches
- NVIDIA Spectrum-5 Ethernet switches
- NVIDIA BlueField-4 DPUs
Related Guides
- ConnectX-8 SuperNIC User Manual
- NVIDIA SHARP Programming Guide
- GPUDirect RDMA and Storage documentation
- IEEE 1588v2 & SyncE deployment guide
- ASAP2 SDN acceleration whitepaper
Want to Know more details about this product
