⚡
Blackwell Ultra Architecture
8x Blackwell Ultra GPUs with 2.3 TB total HBM3e deliver 144 PFLOPS FP4 inference and 72 PFLOPS FP8 training — the most powerful single-node AI server NVIDIA has produced.
UK’s trusted IT infrastructure partner since 2003
144 PFLOPS
FP4 inference per node
2.3TB Memory
Total GPU HBM3e
800G Network
ConnectX-8 per port
10RU Form
Rack-mount chassis
✅ Authorised Reseller
📦 Free UK Delivery
🔒 Full Warranty
📞 24/7 Support
Download Official Datasheet
PDF · Official manufacturer document
NVIDIA DGXSKU: NVIDIA-DGX-B300
NVIDIA DGX B300 — Blackwell Ultra AI Factory Server
The Foundation of the AI Factory Era.
NVIDIA DGX B300, powered by 8x Blackwell Ultra GPUs with 2.3TB total GPU memory, delivers 144 petaFLOPS of FP4 inference — purpose-built for the AI reasoning era and the foundational building block of DGX SuperPOD and DGX BasePOD AI factories.
- 8x NVIDIA Blackwell Ultra GPUs — 2.3 TB total GPU memory (HBM3e)
- 144 PFLOPS FP4 inference / 72 PFLOPS FP8 training per node
- 14.4 TB/s aggregate NVLink bandwidth — all 8 GPUs operate as one unified domain
- Intel Xeon 6776P processors, 112 cores, up to 4 TB system memory
- 8x NVIDIA ConnectX-8 VPI NICs — up to 800 Gb/s InfiniBand/Ethernet per port
- OCP-compliant, MGX-compatible chassis — first in the DGX lineup
- NVIDIA AI Enterprise + NVIDIA Mission Control full software stack included
- 3-year business-standard hardware and software support included
Get Pricing — Speak to Our Team
Request a Quote
Competitive pricing · Response within 4 hours
Key Features
Everything you need in one device
🔗
14.4 TB/s NVLink Interconnect
2x NVLink Switch Systems create a fully connected 8-GPU domain at 14.4 TB/s aggregate — all GPUs communicate as one unified compute entity for trillion-parameter workloads.
🌐
800 Gb/s ConnectX-8 Networking
8x OSFP ports with ConnectX-8 VPI NICs deliver up to 800 Gb/s per port of InfiniBand or Ethernet for scale-out multi-node DGX SuperPOD deployments.
🏭
NVIDIA Mission Control Included
Automates the full AI factory stack — workload scheduling via Run:ai, autonomous cluster recovery, telemetry, provisioning, and diagnostics in one software platform.
🤖
NVIDIA AI Enterprise Suite
NIM microservices for model deployment, optimised frameworks, pretrained models, and DLI training credits for DGX customers — the complete AI software stack.
🏗️
DGX SuperPOD & BasePOD Ready
The foundational node for DGX SuperPOD turnkey AI clusters and DGX BasePOD reference architectures — scale from one node to hundreds seamlessly.
🔄
MGX-Compliant Chassis
First DGX with OCP-compliant and MGX-compatible design — supports both AC and DC-powered deployments across modern data centre environments.
🛠️
3-Year Enterprise Support
3-year business-standard support with 24/7 portal access, live agent support, and NVIDIA DGXpert engineering expertise included with every system.
About the NVIDIA DGX B300
NVIDIA DGX B300 is the most powerful single-node AI server NVIDIA has produced for enterprise data centres. Built on the Blackwell Ultra GPU architecture, DGX B300 contains 8 next-generation GPUs with a combined 2.3 TB of HBM3e memory — enabling enterprises to run AI reasoning workloads, large language model inference, and multi-modal AI training that previously required multiple nodes or cloud resources.
The DGX B300 represents a fundamental shift in enterprise AI infrastructure design. Its OCP-compliant, MGX-compatible chassis allows deployment in both traditional AC-powered and modern DC-powered data centres. At 10 RU with ~14 kW power consumption, DGX B300 delivers hyperscaler-grade AI factory performance in an enterprise-sized footprint with the operational simplicity enterprises require.
Every DGX B300 ships with NVIDIA Mission Control, automating the full operational lifecycle from cluster provisioning to daily workload management, autonomous fault recovery, and utilisation optimisation. Combined with NVIDIA AI Enterprise and NIM microservices, DGX B300 provides a complete develop-to-deploy AI stack validated and supported end-to-end by NVIDIA.
Technical Specifications
📄 Download Datasheet (PDF)Use Cases
Ideal for a wide range of deployments
🧠
AI Reasoning & Inference
Run trillion-parameter reasoning models at production scale with real-time sub-50ms latency — for customer-facing AI and agentic workflow deployments.
🏋️
LLM Fine-Tuning & Training
Fine-tune foundation models on proprietary data and train custom LLMs entirely on a single node — accelerating time-to-production for enterprise AI.
🏭
AI Factory Infrastructure
The foundational compute node of on-premises AI Centres of Excellence — scalable to multi-rack DGX SuperPOD clusters as AI investment grows.
🔬
HPC & Scientific Workloads
Accelerate simulation, genomics, drug discovery, and climate modelling at petaflop scale with Blackwell Ultra FP64 and mixed-precision performance.
FAQ
Frequently Asked Questions
Why Servnet
Why buy from Servnet?
Trusted UK IT reseller since 2003 — supplying businesses of all sizes with genuine, competitively priced technology.
🏆
Authorised UK Reseller
Servnet is an authorised IT reseller with direct access to leading product lines at competitive pricing.
💬
Expert Pre-Sales Advice
Our engineers will help you select the right product and advise on compatibility with your existing infrastructure.
🚚
Fast UK Delivery
Most products ship from UK stock. Express delivery available for urgent deployments.
🔒
Genuine, Warranted Kit
All products are genuine, brand-new, with full manufacturer warranty and RMA support.
📋
Configuration Services
We can pre-configure your hardware before shipping — reducing on-site deployment time and cost.
📞
24/7 Support
Post-sale support from our certified engineers. We're here if anything goes wrong.
