Limited Time Offer: Save $350 on your Ai Cluster → Shop Now

--

Days

--

Hours

--

Mins

--

Secs

Sale Ended

TECHNOLOGY AT A GLANCE

Explore the capabilities

800G OSFP connectivity

Connect compatible InfiniBand and Ethernet fabrics through a single high-density OSFP port.

PCIe 6.0 x16 interface

Bring high-bandwidth network performance into compatible modern server platforms.

RDMA and RoCEv2 acceleration

Help reduce CPU overhead and support efficient data movement across compatible high-performance networks.

Advanced traffic control

Programmable congestion control, quality of service, and telemetry features help manage communication-heavy workloads.

A beige and white industrial PCIe card shown floating against a dark navy background with blue network-like bokeh lines.

800G connectivity for next-generation AI infrastructure

The C8180 Downstream Port SuperNIC brings an 800G network connection into compatible PCIe Gen6 server designs. It is built for high-bandwidth AI, HPC, storage, and cloud environments where efficient communication between accelerators and the network fabric is essential.

TECHNICAL DETAILS

Built into the system

800G XDR network connectivity

The C8180 provides one OSFP network port with up to 800Gb/s of total adapter bandwidth. It is designed for high-density AI, HPC, cloud, and high-performance storage environments that require a fast connection from the server into the network fabric.

InfiniBand and Ethernet flexibility

ConnectX-8 supports compatible InfiniBand and Ethernet deployments. InfiniBand supports speeds up to 800Gb/s, while Ethernet supports compatible 25Gb/s, 50Gb/s, 100Gb/s, 200Gb/s, 400Gb/s, and 800Gb/s connections.

Downstream Port server architecture

This specific DT0 model uses a PCIe Gen6 Downstream Port design for compatible server architectures. It is not the SocketDirect/Multi-Host version and should be selected only for systems designed to support the Downstream Port configuration.

PCIe 6.0 x16 host interface

A PCIe Gen6 x16 interface provides the high-bandwidth server connection needed for modern accelerator, storage, and AI compute platforms. The card uses a half-height, half-length form factor and includes a tall bracket.

RDMA and AI workload acceleration

RDMA, RoCEv2, and GPUDirect RDMA support direct, efficient data movement across compatible systems and fabrics. These capabilities help reduce CPU overhead for distributed AI training, high-performance storage, and communication-intensive HPC workloads.

Traffic management and platform security

Advanced programmable congestion control, quality of service, and telemetry capabilities help manage high-volume network traffic. Secure boot, secure firmware update, flash encryption, and device attestation support platform-level security requirements.