NVIDIA Blackwell Series
HGX B300 Server
The Godì 1.8E2D-NV8 combines eight NVIDIA Blackwell Ultra GPUs with dual Intel® Xeon® 6 Scalable processors to tackle the most demanding AI workloads.
It offers 8 embedded CX8 SuperNIC, 4 PCIe expansion slots, 32 DIMM slots, 12 NVMe drive bays and dual 1GbE LAN.
Purpose-built for LLMs, scientific computing, and AI model training and inference, it is the ideal platform for large enterprises, cloud service providers, research institutions, as well as the financial and automotive sectors.
Image courtesy of NVIDIA. 
NVIDIA Blackwell Ultra
NVIDIA Blackwell breaks new ground with its Transformer Engine, built around a technique called micro-tensor scaling. This approach makes FP4, 4-bit floating point AI, both practical and precise, without sacrificing accuracy.
Memory can now support models twice as large, running twice as fast. NVIDIA's NVFP4 format turns this into a competitive advantage for AI factories, enabling faster training of bigger models at a lower cost.
Solution Rack-Ready
This 48U rack houses 4 high-density nodes, each powered by dual Intel® Xeon® 6th Gen Scalable processors and an NVIDIA HGX-B300-NVL8 baseboard.
Every node supports up to 32 DDR5 DIMM slots and 12 NVMe PCIe5 drives, ensuring both memory bandwidth and storage throughput for the most intensive AI workloads.
Networking is handled by a trio of NVIDIA switches, covering out-of-band management, converged Ethernet, and a dedicated InfiniBand compute fabric, delivering the low-latency, high-bandwidth backbone that large-scale AI training demands.
The rack draws up to 61kW of power and supports both air cooling and an optional rear door heat exchanger for thermally dense deployments.
NVIDIA Networking: 8 embedded CX-8 SuperNIC
The Godì 1.8E2D-NV8 integrates 8 CX8 ports supporting InfiniBand XDR and 800G Ethernet, each directly connected to a B300 SXM module.
This means up to 800G/s of raw throughput, delivered natively, with none of the bottlenecks that come with external adapters.
The benefits go beyond speed. Ultra-low latency keeps AI workloads running at full throttle, while the simplified architecture cuts down on cabling, reduces power consumption, and frees up valuable space inside the server chassis.
Intel® Xeon® 6 Processor Family
The Intel Xeon 6 family combines two complementary architectures to meet modern data center needs.
P-core models (Granite Rapids) deliver maximum performance for AI, HPC, and enterprise workloads, while E-core models (Sierra Forest) provide unmatched efficiency and density for cloud-native and scale-out applications.
With support for DDR5, PCIe 5.0, CXL 2.0, and built-in AI acceleration (Intel AMX), Xeon 6 processors offer the scalability, performance, and efficiency required for next-generation computing.