Unlock AI and HPC At Scale - CN5000 Switch Delivers Unmatched Performance, Scalability, and Efficiency
As AI and HPC workloads grow in complexity high-bandwidth networking becomes essential. The Cornelis CN5000 Omni-Path Switch is designed to meet the extreme scalability and performance demands of modern compute environments, ensuring rapid results and maximizing infrastructure investments.
The CN5000 Switch is engineered to scale seamlessly across thousands of nodes, leveraging advanced network routing, intelligent traffic flow management, and zero-latency packet integrity protection. These capabilities enhance workload efficiency by minimizing congestion, preserving data accuracy, and optimizing GPU utilization—resulting in improved performance and lower total cost of ownership (TCO).
Delivering up to 400Gbps of bidirectional bandwidth, the CN5000 Switch is designed to support the most demanding AI training and inference tasks. It ensures ultra-low latency and eliminates bottlenecks, accelerating time-to-results for generative and agentic AI applications. With unparalleled robustness and efficiency, it remains the ideal choice for parallelized compute environments and HPC workloads.
Consistent, reliable network performance is critical for AI and HPC applications. The CN5000 Switch incorporates advanced error correction and dynamic network function placement, enhancing overall system efficiency and adaptability to evolving workloads.
Equipped with cutting-edge telemetry capabilities, the CN5000 Switch provides real-time network insights and fine-grained visibility to optimize workload performance. By continuously analyzing traffic patterns, identifying congestion points, and dynamically adjusting network flows, it ensures smooth, high-efficiency operation—even in the most complex compute environments. With industry-leading telemetry and precision network intelligence, CN5000 enables greater control, efficiency, and performance across any scale. The Cornelis CN5000 product family includes the Omni-Path Switch, Director Class Switch, Adapter, Copper and Optical cables, and open-source Host & Management Software—offering flexible, high-performance networking solutions for diverse infrastructure needs.
48 ports, each providing 400G performance
32 ports independently support port subdivision to 2 x 200G each
Hot swappable field replaceable units (FRUs): fan and power supply
38.4 Tbps aggregate bandwidth
Optimization for message rate and latency
Virtual lanes (VLs): Configurable from one to four VLs plus one management VL
Configurable MTU size of 2, 4, 8, or 10KB
Maximum multicast table size: 16,384 entries
Maximum unicast table size: 49,152 entries (linear LIDs)
Maximum unicast table size: 330,000 entries (hierarchical LIDs)
Fine grain, telemetry-based dynamic adaptive routing (FGAR)
Integrated OpenBMC-based management
Redfish protocol and data model/schema support
In-band and out-of-band management options
Command line interface through 10/100/1000 BASE-T Ethernet
Serial console through USB A serial port
Network Time Protocol (NTP) and LDAP
Static Dispersive Routing (SDR)
FGAR or SDR selectable per packet
Innovative traffic categorization to control incast-prone applications
QSFP112 Quad Small Form Factor Pluggable cabling
Low-Latency bit error recovery and optional correction
Enhanced congestion control that can throttle adapter injection
Rich Quality of Service (QoS) controls
Backwards compatible with previous generation Omni-Path 100G fabric
Open BMC-based chassis management
Security (Secure key EEPROM, Secure Boot)
Options for air and liquid cooling
| Bandwidth | 400G |
| Fabric Ports | 48 x 400G QFSP112 ports |
| Chassis Dimensions | 8.4 x 22.0 x 31.9 in. (4.40 x 43.79 x 66.30 cm) (H x W x D) |
| Weight | Air Cooled Liquid Cooled |
| Power Consumption (Typical/Max) | Cable Configuration:
|
| Input Range | 200-277 VAC, 50-60 Hz |
| Cooling Options | Air Cooling Full Liquid Cooling |
US/Canada | NRTL: UL 62368-1, CSA 22.2 No. 62368-1 |
Europe | EN62368-1 CB Scheme: IEC 62368-1 |
Temperature | Operating: 10° to 35° C (up to 900m above sea level. Temperature is derated 1C per 175m above 900m) Storage: -40° to 70° C |
Humidity | Operating: 5% to 90% non-condensing Storage: 5% to 95% non-condensing |
Altitude | Operating: 0-3,000m (Temperature derating 1˚C/175m above 900m) Non-operating: 0-12,000m |
US/Canada | FCC Part 15, Subpart B, Class A CAN ICES-3 (A)/NMB-3(A) Issue 7 |
Europe/International | EN55032 (CISPR32) EN55035 (CISPR35) EN61000-3-2 EN61000-3-3 |
Japan | VCCI, Class A |
New Zealand/ | AS/NZS CISPR 32, Class A |
Korea | Emissions: KS C 9832, Class A Immunity: KS C 9835 |
Taiwan | BSMI (CNS 15936), Class A |
Country of Origin | Malaysia |
From training large language models and powering agentic AI frameworks, to real-time fraud detection in finance and computer vision in autonomous systems - CN5000 Omni-Path AI solutions deliver the high-performance, lossless networking foundation that next-generation AI demands.
Push the boundaries of scientific discovery and engineering innovation with CN5000 Omni-Path-based HPC solutions. From advancing climate modeling and seismic analysis, to accelerating simulation and design in aerospace and automotive, and enabling complex genomic research in life sciences, CN5000 Omni-Path solutions power breakthrough results—faster, smarter, and at scale.
