EMARQUE AI HGX Rubin RGE8 AI Server

The EMARQUE AI HGX Rubin RGE8 is EMARQUE's NVIDIA HGX Rubin NVL8 server — a liquid-cooled 8-GPU frontier-AI system with 8× NVIDIA Rubin GPUs on the NVIDIA HGX Rubin baseboard, joined by sixth-generation NVLink and NVLink Switch at 3.6 TB/s GPU-to-GPU, and driven by dual Intel Xeon 6 processors. It's built for the next generation of frontier LLM training, reasoning models and the highest-throughput production inference. EMARQUE configures, integrates and delivers it in Malaysia — cooling configuration, networking, memory and storage to your workload; EMARQUE scopes, integrates and supports it locally.

Each system ships as a complete liquid-cooled server: the HGX Rubin NVL8 baseboard with 8× Rubin GPUs and HBM4, dual Intel Xeon 6, NVIDIA ConnectX-9 SuperNICs, Gen5 E1.S NVMe storage, and direct-liquid CPU + GPU cooling with leak detection. EMARQUE AI Solutions are available on request — deployment scoping, rack and coolant-distribution-unit (CDU) planning, InfiniBand / Ethernet fabric design, NCCL and MPI tuning, multi-server cluster setup, orchestration, private AI stack deployment, team training. Final configuration confirmed on enquiry.

The EMARQUE AI HGX Rubin RGE8 follows the NVIDIA HGX Rubin NVL8 platform specification:

• GPUs: 8× NVIDIA Rubin on the NVIDIA HGX Rubin NVL8 baseboard; sixth-generation NVLink + NVLink Switch, 3.6 TB/s GPU-to-GPU bandwidth.
• GPU memory: next-generation HBM4 — up to 288 GB per Rubin GPU (up to ~2.3 TB across the 8-GPU baseboard).
• AI compute: up to ~400 PFLOPS NVFP4 inference (8× Rubin at 50 PFLOPS each); ~280 PFLOPS NVFP4 training — the Rubin-generation step beyond Blackwell Ultra.
• CPU: dual Intel Xeon 6 (6700 / 6500-series).
• System memory: 8-channel DDR5 RDIMM / MRDIMM per CPU, up to 32 DIMMs.
• GPU networking: 8× 800 Gb/s OSFP (InfiniBand or Ethernet) via NVIDIA ConnectX-9 SuperNICs; NVIDIA BlueField-4 DPU (FHHL PCIe Gen6 x16).
• Storage: 8× E1.S Gen5 NVMe hot-swap + 2× M.2 (PCIe Gen6).
• Cooling: direct liquid cooling (CPU + GPU) with leak detection — hybrid direct-to-chip (liquid GPU baseboard + air-cooled CPUs) or 100% liquid-cooled configurations; 54 V DC bus-bar power delivery.
• GPU power: approximately 2.3 kW at the GPU baseboard.
• Form factor: liquid-cooled rackmount (2U–3U per cooling configuration), modular with independent CPU and GPU sleds and tool-less access.
• Software: NVIDIA AI Enterprise stack; DGX-OS-class environment scoped per deployment.

Exact configuration (Xeon SKU, memory, NVMe capacity, networking, cooling) and final pricing confirmed on enquiry — email business@emarque.co.

NVIDIA HGX Rubin NVL8 is NVIDIA's Rubin-generation baseboard platform for frontier AI — the successor to HGX B300. It mounts 8× Rubin GPUs with sixth-generation NVLink and NVLink Switch (3.6 TB/s GPU-to-GPU), giving every GPU high-bandwidth access to every other GPU in the box, updated to Rubin silicon with next-generation HBM4 memory and a major step up in NVFP4 AI throughput. It sits within NVIDIA's wider Rubin platform, whose rack-scale flagship is the Vera Rubin NVL72.

The EMARQUE AI HGX Rubin RGE8 is one of the HGX Rubin NVL8 server platforms in EMARQUE's lineup. HGX Rubin NVL8 is the right platform for organisations training and serving next-generation frontier models on-prem at the highest throughput, building AI factories at multi-server scale, and meeting sovereign-AI requirements where data must stay in country. EMARQUE is among the first solution providers in Malaysia taking pre-orders for the Rubin generation.

The EMARQUE AI HGX Rubin RGE8 ships with a 3-year hardware warranty covering hardware faults. EMARQUE's built-in role is Malaysia-based local handling:

• Single point of contact — pre-sale, post-sale and warranty routed through the same EMARQUE team.
• RMA routing — if a hardware fault appears, EMARQUE routes the warranty claim on your behalf.

For deeper post-sale needs — pre-deployment burn-in, liquid-cooling and CDU commissioning, OS / driver / NCCL tuning, InfiniBand or Ethernet fabric design, multi-server cluster setup, orchestration, private AI stack deployment, team training — solutions are available from the EMARQUE AI team on request. Scope and charges quoted on enquiry. Software / OS / driver / setup support is scoped to the 3-year warranty window.

Step up vs the EMARQUE AI HGX B300 HGE8: the HGX B300 HGE8 is the current Blackwell Ultra generation — shipping today, 8× B300 with 2.1 TB HBM3E, air-cooled. The HGX Rubin RGE8 is the next-generation Rubin step: 8× Rubin GPUs, HBM4, higher NVFP4 throughput, liquid-cooled. Choose RGE8 when your timeline aligns with the Rubin generation and you want next-generation hardware.

Step down vs Vera Rubin NVL72: Vera Rubin NVL72 is NVIDIA's rack-scale Rubin flagship — 72 Rubin GPUs + 36 Vera CPUs, 3.6 exaFLOPS NVFP4 in one rack. The HGX Rubin RGE8 is the 8-GPU server tier of the same Rubin generation — the right choice when a single high-density server (or a small multi-server cluster) fits, rather than a full rack.

EMARQUE AI sizes the right tier — RGE8 server, HGE8 server, or NVL72 rack — against your workload, timeline and datacentre (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 on the Rubin generation.
• Highest-throughput inference — multi-tenant serving with NIM microservices, vLLM, TensorRT-LLM.
• Reasoning and agentic AI — long-context inference and autonomous workflows with secure tool-calling.
• Retrieval-augmented generation at organisation scale — embedding pipelines, vector search, document AI.
• Computer vision and multimodal — VLM training and inference, video pipelines.
• RAPIDS for massive-dataset data science.

Real-world: private frontier-model fine-tuning on enterprise data, organisation-wide internal LLM serving, sovereign-AI deployments where data must stay on-prem.

EMARQUE AI Servers ship with the full NVIDIA AI Enterprise stack — CUDA, container runtimes, NIM microservices for production inference, AI Workbench, and pre-built framework images (PyTorch, vLLM, RAPIDS, NeMo). The HGX Rubin RGE8 runs a DGX-OS-class environment with Base Command for multi-tenant orchestration, scoped per deployment.

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 — chat UI, vector search, document ingestion, model serving, all on-prem.

The NVIDIA HGX Rubin NVL8 platform is entering the market as NVIDIA's Rubin generation ramps into production. Pre-ordering the EMARQUE AI HGX Rubin RGE8:

• Locks your allocation for delivery in Malaysia.
• Final pricing confirmed on enquiry — pricing varies by CPU configuration, cooling configuration (hybrid vs full liquid), system memory, NVMe capacity, networking, and EMARQUE AI services scope. Listed price is an indicative estimate.
• Delivery window aligned with NVIDIA's Rubin ramp and the Malaysia shipping schedule — typical lead time confirmed at order.
• Direct access to the EMARQUE AI team during pre-order for deployment scoping (rack, CDU / liquid cooling, 3-phase power, InfiniBand or Ethernet fabric, multi-server cluster 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 NVIDIA HGX Rubin NVL8 servers. What sets EMARQUE apart on this tier is the depth of solutions available from the EMARQUE AI team:

• Liquid-cooling deployment — CDU sizing, coolant loop and leak-detection commissioning, rack and 54 V bus-bar power, datacentre integration for direct-liquid-cooled AI. Scope and charges quoted on enquiry.
• Deployment scoping — InfiniBand vs Ethernet fabric decision, multi-server cluster topology, NCCL and MPI tuning.
• Tier sizing — honest assessment of whether the HGX Rubin RGE8, the current-generation HGX B300 HGE8, or a Vera Rubin NVL72 rack fits your workload, timeline and datacentre.
• Orchestration — Base Command / multi-tenant orchestration setup, model serving, internal AI platform layer.
• Local point of contact — pre-sale sizing, post-sale technical questions, warranty handling — same Malaysian team, no offshore queue.

For buyers comparing HGX Rubin providers, the practical question is who can integrate the server, the liquid cooling, 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.