An NVIDIA HGX B300 8-GPU system in 4U with direct liquid cooling, sold and supported by Cisco and managed through Cisco management platforms for AI training and high-density inference.
An NVIDIA HGX B300 8-GPU system in 4U with direct liquid cooling, sold and supported by Cisco and managed through Cisco management platforms for AI training and high-density inference.
This server brings NVLink-connected training and highdensity inference capacity to the Cisco AI compute portfolio. The underlying hardware is the Supermicro SuperServer SYS-422GS-NB3RT-LCC, a 4U liquidcooled NVIDIA HGX B300 8-GPU system. Cisco handles the selling motion, support through Cisco CX Services. Supermicro handles liquid cooling maintenance and support services. Lifecycle management through Cisco Intersight or Hyperfabric. Supermicro builds the hardware. Customers get Supermicro’s liquid-cooled GPU engineering and Cisco’s support and operations model in one orderable product.
The system pairs two Intel Xeon processors with an NVIDIA HGX B300 8-GPU baseboard connected over fifth-generation NVLink at 1.8 TB/s per GPU, with 2.3 TB of HBM3e memory across the baseboard. Direct-to-chip liquid cooling moves GPU and CPU heat into facility water and fits the 8-GPU system into 4U instead of the 8U an air-cooled equivalent needs. Target workloads include large-model training, fine-tuning, and high-throughput inference where eight NVLink-connected GPUs per node are required.
Table 1 summarizes the primary features and benefits of the platform
| Feature | Benefit |
|---|---|
| NVIDIA HGX B300 8-GPU baseboard with NVLink 5 at 1.8 TB/s | Eight GPUs work as one coherent pool for training and long-context inference |
| 2.3 TB HBM3e GPU memory per system | Large models and KV caches stay resident on fewer nodes |
| Direct-to-chip liquid cooling | GPU and CPU heat leaves through facility water; lower fan power, noise, and cooling energy |
| 4U form factor | Twice the GPU density per rack compared with 8U air-cooled HGX B300 systems |
| Dual Intel Xeon 6768P processors | Keeps data pipelines fed, with up to 4 TB of DDR5-6400 in the Cisco configuration |
| 8x NVIDIA ConnectX-8 SuperNICs at 800 Gb/s | One NIC per GPU for non-blocking east-west scale-out |
| NVIDIA BlueField-3 B3240 DPU | Offloads north-south networking and storage from the host CPUs |
| Cisco Intersight management (future) | One operations model across the Cisco compute estate |
| Supermicro Services with liquid cooling maintenance | Dedicated support contracts covering the server and the liquid loop |
Eight Blackwell Ultra GPUs, one NVLink domain, four rack units.
Air-cooled HGX B300 servers need 8U and move heat with fans. This system puts the same eight-GPU NVLink domain in 4U by running direct-to-chip cold plates across the GPUs and CPUs. Most system heat leaves through the liquid loop rather than the data center air handlers, which doubles GPU density per rack and reduces cooling energy and noise at the same time.
Scale-out is sized to match: Eight NVIDIA ConnectX-8 SuperNICs give each GPU its own 800 Gb/s east-west port, and an NVIDIA BlueField-3 B3240 DPU carries north-south and storage traffic. In the Cisco offer the platform is deployed with 415 V rack PDUs and an in-rack or in-row coolant distribution unit, so the liquid infrastructure is part of the order, not a separate project.
Cisco offers this Supermicro platform in a defined configuration, listed in Table 2.
| Item | Selected configuration |
|---|---|
| PID | SYS422GSNB3RTLCC1 SYS422GSNB3RTLCC2 |
| Platform | HGX 4U rackmount liquid-cooled platform |
| Dimmensions | 17.6 in W x 6.85 in H x 39.03 in D |
| GPU | 1x NVIDIA HGX B300 8-GPU baseboard |
| CPU | 2x Intel Xeon 6768P |
| Memory | 16x or 32x 128 GB DDR5-6400 RDIMM (2 TB or 4 TB total) |
| M.2 boot drives | 2x1.9TB NVMe PCIe4x4 M.2 22x110mm (1DWPD) |
| NVMe storage | Qty 8 3.84TB E1.S drives |
| East-west NICs | 8x NVIDIA ConnectX-8 SuperNICs |
| North-south NICs | 1x NVIDIA BlueField-3 B3240 DPU |
| Software management | Cisco Nexus One or Cisco Intersight |
| Rack PDU | 415 V |
| CDU | In-rack or in-row/sidecar |
| Support | Cisco CX Services plus Supermicro liquid cooling maintenance |
| Availability | October 2026 |
This is a hardware product; no platform license is required. Cisco Intersight management requires an Intersight license, listed in Table 3.
| License | Description |
|---|---|
| Cisco Intersight | License tier appropriate to the deployment, see Cisco sales specialist |
Information about Cisco’s Environmental, Social and Governance (ESG) initiatives and performance is provided in Cisco’s CSR and sustainability reporting.
| Sustainability Topic | Reference | |
|---|---|---|
| General | Information on product-material-content laws and regulations | Materials |
| Information on electronic waste laws and regulations, including our products, batteries and packaging | WEEE Compliance | |
| Information on product takeback and resuse program | Cisco Takeback and Reuse Program | |
| Sustainability Inquiries | Contact: csr_inquiries@cisco.com | |
| Material | Product packaging weight and materials | Contact: environment@cisco.com |
Table 5 reflects the Supermicro data sheet for the SYS-422GS-NB3RT-LCC. Product specifications may change without notice.
| Category | Specification |
|---|---|
| Form factor | 4U rackmount, liquid cooled; enclosure 448 x 174 x 896.8 mm (17.6 in W x 6.85 in H) |
| GPU | NVIDIA HGX B300 8-GPU baseboard; 8x SXM modules, 2.3 TB total HBM3e |
| GPU interconnect | NVIDIA NVLink 5 with NVSwitch, 1.8 TB/s GPU-to-GPU |
| Processor | Dual socket Intel Xeon 6700-series with P-cores, up to 86 cores/172 threads per CPU, up to 350 W, up to 336 MB cache |
| System memory | 32 DIMM slots; up to 8 TB DDR5-5200 (2DPC) or 4 TB DDR5-6400 (1DPC) |
| Drive bays | 8 front hot-swap E1.S NVMe bays; 2x M.2 NVMe with hardware RAID (Broadcom 3808N) |
| Expansion slots | 2x PCIe 5.0 x16 FHHL double-width; 2x PCIe 5.0 x16 LP |
| Networking | 8x 800 Gb/s OSFP ports (NVIDIA ConnectX-8 SuperNICs); 2x RJ45 10GbE (Intel X710) |
| Cooling | Direct-to-chip (D2C) cold plates plus 4x 80 mm heavy duty fans |
| Power | 4x 6600 W redundant (2+2) Titanium Level |
| BIOS | AMI 64 MB UEFI |
| System management | Cisco Intersight (Cisco offer) |
| Operating environment | 10 to 35 C operating |
| Weight | Net 84.6 kg (186.5 lb); gross 100.7 kg (222 lb) |
| Motherboard/chassis | Super X14DBG-LC/CSE-GP405LTS-R000NP |
Table 6 lists what must be in place before installation.
| Requirement | Description |
|---|---|
| Rack integration | Standard 19-inch rack, 4U, 896.8 mm chassis depth; verify rail and manifold clearance |
| Facility cooling | Coolant distribution unit, in-rack or in-row/sidecar; facility water loop for the direct-to-chip cold plates |
| Power | 4x 6600 W inputs; 415 V rack PDU |
| Operating temperature | 10 to 35 C (50 to 95 F) |
| Management | Cisco Intersight account and claimed device connectivity |
The base platform is a Supermicro SKU offered through Cisco. Cisco ordering PIDs for the configured system are in progress. To order, contact your Cisco account representative or visit the Cisco Ordering Home Page.
| Part number | Product description |
|---|---|
| SRS-48UDLC-B300-M0 | Supermicro 48U DLC Rack, HGX B300 NVL8 422GS, no In-Rack CDU |
| SRS-48UDLC-B300-M1 | Supermicro 48U DLC Rack, HGX B300 NVL8 422GS, with In-Rack CDU |
| LCS-SCDU-1K3LR001 | 1.8MW In-Row CDU, L2L, N+1 Pumps |
| LCM-SMLCPM3-3Y | Liquid Cooling Maintenance Service, SMLCPM3, 3-Year |
| LCS-SCDU-200AR001 | 200kW L2A Sidecar CDU, N+1 Pumps |
| SVC-NVSTDSUP-3Y | NVIDIA Enterprise Support Entitlement, 3Yr |
Warranty terms for this Cisco-offered Supermicro platform are standard Supermicro warranty terms. Support is delivered through Cisco CX Services under the applicable service contract, contact your Cisco sales specialist for more details.
Cisco CX Services covers this platform from planning through deployment and ongoing operations, including liquid cooling maintenance, with a single Cisco support contract even though the hardware is Supermicro-built. For more information, visit https://www.cisco.com/go/services.
Cisco Capital makes it easier to get the right technology to achieve your objectives, enable business transformation and help you stay competitive. We can help you reduce the total cost of ownership, conserve capital, and accelerate growth. In more than 100 countries, our flexible payment solutions can help you acquire hardware, software, services and complementary third-party equipment in easy, predictable payments. Learn more.
For more on Cisco AI infrastructure, visit https://www.cisco.com/go/ai. Full platform specifications are available on the Supermicro product page for the SYS-422GS-NB3RT-LCC.
Learn more about Cisco and Supermicro partnership:
https://www.cisco.com/site/us/en/solutions/global-partners/supermicro/index.html.