DC AI Multi-GPU Montjuïc – AMD Ryzen Threadripper 9960X | RTX 5090 32GB x2 | 128GB DDR5 | 2TB Gen5 NVMe | Ubuntu AI Workstation

Product Description

The DC AI Multi-GPU – Montjuïc is where AI workstations evolve into true high-performance compute systems. Designed for advanced developers, AI engineers, and teams working with demanding models and large-scale workflows, this system delivers parallel GPU power, massive memory bandwidth, and workstation-class stability.

Inspired by the scale and spectacle of the Magic Fountain of Montjuïc, this machine represents synchronised power — multiple forces working together in perfect harmony to deliver results at a level single-GPU systems simply cannot match.

⚙️ Threadripper-Class Processing Power

At the core is the AMD Ryzen Threadripper 9960X, a high-core-count workstation processor built for extreme multitasking, data processing, and parallel workloads.

This enables:

• Faster data preprocessing and pipeline execution
• Seamless handling of multiple AI environments
• High-throughput development and simulation workflows

This is the foundation required for true multi-GPU computing.

🔥 Dual GPU Performance – A New Level of Capability

Equipped with 2× NVIDIA RTX 5090 32GB GPUs, this system unlocks a completely new tier of AI performance:

• Run multiple models simultaneously
• Accelerate training and inference workloads
• Distribute workloads across GPUs for efficiency
• Handle larger and more complex AI pipelines

This is where performance stops being linear — and starts to scale.

🧠 Massive System Memory for Heavy Workloads

With 128GB of DDR5 RAM, the Montjuïc is built to handle large datasets, multiple containers, and memory-intensive processes without compromise. Perfect for developers working across multiple environments or managing large-scale AI projects.

🚀 Ultra-Fast Gen5 Storage

A 2TB Crucial P510 Gen5 NVMe SSD delivers up to 10,000MB/s read speeds, ensuring datasets, models, and applications load instantly. Whether you're working with massive training data or large AI models, your workflow remains fast and uninterrupted.

🧊 Workstation Cooling for Sustained Load

The ARCTIC Liquid Freezer WS360-SP6 is specifically designed for workstation-class CPUs, ensuring the Threadripper 9960X maintains peak performance under sustained workloads. Combined with the high-airflow Phanteks Enthoo Pro II Server Edition chassis, this system is built for continuous operation under heavy load.

🧩 Built for Multi-GPU Infrastructure

Housed on the ASUS PRO WS TRX50-SAGE WIFI motherboard, this platform is engineered for stability, expansion, and professional workloads — providing the bandwidth and reliability required for dual-GPU AI systems.

🌐 Connected & AI-Ready

With Gigabit Ethernet + WiFi connectivity, the system integrates seamlessly into local or hybrid environments. Pre-installed with Ubuntu, it is ready for AI frameworks such as PyTorch, TensorFlow, CUDA, and distributed workloads straight out of the box.

🔌 High-Capacity, Ultra-Stable Power

The Corsair HX1500i (2025) PSU delivers ultra-efficient, low-noise power for dual GPU operation and sustained compute workloads — ensuring stability even under peak demand.

🚀 Scale Your AI Performance

The DC AI Multi-GPU – Montjuïc is built for those who need more than raw power — it’s designed for scalability. With dual GPUs working in parallel and a workstation-class platform underneath, this system enables faster iteration, larger workloads, and more efficient AI development.

Inspired by the Magic Fountain of Montjuïc, it represents synchronised performance at scale — where multiple systems work as one to deliver something greater.

🧠 Perfect For

• Multi-GPU AI development & training
• Running multiple models simultaneously
• Faster inference and iteration cycles
• Advanced machine learning pipelines
• AI engineers, studios, and development teams

💡 Part of the DC AI Fountain Series

From Trevi to Swarovski, capability grows — but Montjuïc is where performance scales. This is the step into true high-performance AI computing, where systems don’t just run models — they accelerate them.

Specifications

*** This specification may be subject to change depending on availability. Only like for like components used. You'll be contacted before. Images for illustrative purposes only. ***

FAQ's

Product Gallery