NVIDIA DGX B300 AI Server

The NVIDIA DGX B300 is NVIDIA's own first-party AI system — the fully-integrated, NVIDIA-built, NVIDIA-supported appliance built on Blackwell Ultra, and the foundation of NVIDIA DGX BasePOD and DGX SuperPOD. Each system ships as a complete turnkey unit: a 10U chassis with 8× NVIDIA B300 Blackwell Ultra GPUs, dual Intel Xeon 6776P CPUs, NVIDIA ConnectX-8 SuperNICs and BlueField-3 DPUs, and the complete DGX software stack preinstalled — DGX OS, NVIDIA AI Enterprise, NVIDIA Base Command and NVIDIA Mission Control.

DGX B300 is the premium NVIDIA-turnkey tier — one tested, optimised hardware-and-software platform backed by NVIDIA Enterprise Services. EMARQUE AI Solutions are available on request — deployment scoping, rack and 3-phase PDU sizing, fabric design, multi-node SuperPOD planning, private AI stack deployment, team training. Final configuration confirmed on enquiry.

The NVIDIA DGX B300 follows NVIDIA's published specification:

• GPUs: 8× NVIDIA Blackwell Ultra (SXM), 288 GB HBM3E per GPU.
• GPU memory: 2.1 TB total HBM3E with 64 TB/s aggregate memory bandwidth.
• AI compute: 144 PFLOPS sparse FP4 (108 PFLOPS dense) for inference; 72 PFLOPS FP8 for AI training — 2× the attention performance and 1.5× the dense FP4 of DGX B200.
• CPU: dual Intel Xeon Platinum 6776P.
• System memory: 2 TB DDR5 standard, configurable up to 4 TB.
• Storage: 2× 1.92 TB NVMe M.2 (OS, self-encrypting) + 8× 3.84 TB NVMe E1.S (data / cache, self-encrypting).
• GPU interconnect: 5th-generation NVIDIA NVLink + NVLink Switch — 14.4 TB/s all-to-all GPU-to-GPU bandwidth.
• Networking: 8× NVIDIA ConnectX-8 SuperNICs (up to 800 Gb/s InfiniBand or Ethernet, connected directly to the GPUs over PCIe Gen6) + 2× NVIDIA BlueField-3 DPUs (up to 400 Gb/s).
• Power: ~14.5 kW max; 12× 3.2 kW power supplies, N+N redundant; busbar or traditional PDU.
• Form factor & cooling: 10U rackmount, air-cooled; MGX and traditional-enterprise-rack compatible.
• Software / OS: DGX OS 7 (Ubuntu 24.04 LTS base) + NVIDIA AI Enterprise + NVIDIA Base Command + NVIDIA Mission Control (Base Command + Run:AI) + NGC catalogue.

Exact configuration and final pricing confirmed on enquiry — email business@emarque.co.

NVIDIA DGX B300 is NVIDIA's own universal AI system — built, integrated and supported by NVIDIA, the successor to DGX B200 / DGX H200. It packs 72 petaFLOPS of AI training performance into a 10U form factor with 8× Blackwell Ultra GPUs, 2.1 TB HBM3E and 64 TB/s of memory bandwidth, all unified by 5th-gen NVLink. It is the foundation of NVIDIA DGX BasePOD and DGX SuperPOD — the building block enterprises use to scale from one system to turnkey supercomputing clusters.

DGX B300 is the right platform for organisations that want NVIDIA's own engineering, the complete DGX software stack, and NVIDIA Enterprise Services as a single tested standard — frontier LLM training, the highest-throughput inference, AI factories, and sovereign-AI deployments. EMARQUE is among the first solution providers in Malaysia taking pre-orders for this tier.

The NVIDIA DGX B300 is backed by NVIDIA Enterprise Services — NVIDIA's own hardware and software support, education and professional services for DGX infrastructure, available in 3-, 4- or 5-year terms. This is NVIDIA-direct support for NVIDIA's own system. EMARQUE's built-in role is Malaysia-based local handling:

• Single point of contact — pre-sale, post-sale and the NVIDIA support relationship routed through the same EMARQUE team.
• Local coordination — EMARQUE coordinates the NVIDIA Enterprise Services engagement on your behalf.

For deeper deployment needs — rack and fabric design, multi-node SuperPOD setup, Base Command / Mission Control orchestration, private AI stack deployment, team training — solutions are available from the EMARQUE AI team on request. Scope and charges quoted on enquiry.

Both are 8× B300 Blackwell Ultra with 2.1 TB HBM3E. The difference is the integration tier:

NVIDIA DGX B300 — NVIDIA's own turnkey system. 10U, fixed dual Intel Xeon 6776P, the complete DGX software stack (DGX OS + Mission Control + Base Command), NVIDIA Enterprise Services, and the validated foundation of DGX SuperPOD. Choose it when you want NVIDIA's own engineering, software and support as one tested standard, or when you're scaling into a SuperPOD.

GIGABYTE G894 (HGX B300) — GIGABYTE's partner-built HGX B300 server. 8U, your choice of AMD EPYC or Intel Xeon CPU, more configuration flexibility and a lower total cost of ownership, with GIGABYTE warranty + EMARQUE local support. Choose it when configurability and TCO matter more than the full NVIDIA-turnkey stack.

Below both sit the EMARQUE-built RTX PRO 6000 Blackwell servers (GDDR7, PCIe, no NVLink baseboard) for inference-density workloads. EMARQUE AI sizes the right tier against your workload (email business@emarque.co).

Production AI workloads — the things that exceed a single workstation:

• Frontier LLM training and fine-tuning — pre-training, SFT, RLHF, DPO at production scale.
• Highest-throughput inference — multi-tenant serving with NIM microservices, vLLM, TensorRT-LLM.
• Retrieval-augmented generation at organisation scale — embedding pipelines, vector search, document AI.
• Agentic AI and reasoning models with secure tool-calling.
• Computer vision, multimodal, physical AI training.
• RAPIDS for massive-dataset data science.

As the foundation of DGX SuperPOD, DGX B300 is also built to compose into multi-node clusters for the largest training runs.

DGX B300 ships with the complete NVIDIA DGX software stack: DGX OS (Ubuntu / RHEL base with optimised GPU and networking drivers, Docker, container and system management), the full NVIDIA AI Enterprise suite (CUDA, NIM microservices, framework images — PyTorch, vLLM, RAPIDS, NeMo), NVIDIA Base Command and NVIDIA Mission Control (Base Command + Run:AI for orchestration, Slurm and Jupyter), and access to the NGC catalogue of tuned containers, pre-trained models and workflows.

EMARQUE AI Solutions on request can layer in a ready-to-use private LLM, RAG, embeddings and agentic-AI stack pre-configured for your data and use case.

The NVIDIA DGX B300 is shipping now. Pre-ordering with EMARQUE:

• Locks your NVIDIA DGX allocation for delivery in Malaysia.
• Final pricing confirmed on enquiry — pricing depends on the support term (3 / 4 / 5 years), NVIDIA AI Enterprise term, system configuration and EMARQUE AI services scope. Listed price is an indicative estimate.
• Delivery window aligned with NVIDIA's Malaysia allocation schedule — typical lead time confirmed at order.
• Direct access to the EMARQUE AI team during pre-order for deployment scoping (rack, 3-phase PDU, fabric, SuperPOD planning, software stack).

Click ENQUIRE NOW above or email business@emarque.co.

EMARQUE is among the first solution providers in Malaysia taking pre-orders for the NVIDIA DGX B300. What sets EMARQUE apart is the depth of solutions available from the EMARQUE AI team:

• Deployment scoping — rack and 3-phase PDU sizing, datacentre cooling, fabric design, multi-node DGX SuperPOD topology. Scope and charges quoted on enquiry.
• Tier sizing — honest assessment of whether the NVIDIA DGX B300 turnkey tier, the GIGABYTE G894 HGX B300, or an EMARQUE-built RTX PRO 6000 server fits your workload, rack and budget.
• Orchestration — Base Command / Mission Control setup, model serving, internal AI platform layer.
• Local point of contact — pre-sale sizing, post-sale coordination of NVIDIA Enterprise Services — same Malaysian team, no offshore queue.
• EMARQUE AI stack + assisted setup — pre-configured private stack (chat UI, RAG with vector search, document ingestion, model serving) tuned for your data.

For buyers comparing DGX providers, the practical question is who can integrate the system, the rack, the fabric and the AI stack — in country, with one point of contact. Email business@emarque.co for sizing.

Yes. EMARQUE AI is our dedicated AI compute division — full NVIDIA AI hardware stack and on-prem deployment for Malaysian businesses, research labs, and enterprises: NVIDIA Personal AI (DGX Spark, 2× Spark stacking), AI workstations (RTX PRO 6000 Blackwell and others), DGX Station GB300, multi-GPU AI servers, plus full integration support. Visit EMARQUE.AI or contact business@emarque.co.