Vietnam Builds $30K AI Rig with Seven RTX 5090 GPUs

Hanoi, Vietnam – In a testament to the growing demand for accessible AI computing power, Vietnamese retailer Nguyencongpc has assembled a high-performance AI system featuring seven NVIDIA RTX 5090 graphics cards. This custom-built machine, designed for AI developers and researchers, highlights the innovative ways users are leveraging consumer-grade hardware to tackle demanding AI workloads.

According to a report by ILeakVN, the system utilizes an open-air frame and is powered by multiple Super Flower Leadex 2000W power supplies. The estimated cost of each RTX 5090 card ranges from $3,500 to $4,000, pushing the total hardware cost of the system above $30,000.

Despite the scarcity of the RTX 5090, its impressive AI processing capabilities and lower price point compared to NVIDIA’s Blackwell B100/B200/B300 series GPUs make it an attractive option for AI acceleration.

The system connects the seven RTX 5090 cards via PCIe risers, creating a unified 224GB memory pool. This configuration requires model parallelism techniques to enable multi-card collaboration, allowing the system to handle complex AI models.

This approach offers a compelling alternative to NVIDIA’s professional-grade RTX Pro 6000, which features 96GB of memory but carries a significantly higher price tag of $8,000 to $9,000 per card.

The demand for RTX 5090 cards for AI applications is contributing to supply constraints in the consumer market. NVIDIA’s prioritization of AI accelerators, which offer higher profit margins, has led to limited availability of consumer-grade cards. Gamers seeking high-end configurations may face inflated prices and limited options, as evidenced by the emergence of modified RTX 4090 turbine cards with 48GB of memory.

For AI developers with budget constraints, the RTX 5090 multi-card solution presents a cost-effective alternative to the Blackwell B100 accelerator. However, it’s important to note that professional-grade cards with larger memory capacities remain the preferred choice for processing massive AI models with billions of parameters.

The Implications

The creation of this seven-card RTX 5090 AI system underscores several key trends:

Growing Demand for Accessible AI: The need for affordable AI computing power is driving innovation in hardware configurations.
Resourcefulness of AI Developers: Users are finding creative ways to leverage consumer-grade hardware to meet their AI processing needs.
Supply Chain Dynamics: NVIDIA’s prioritization of AI accelerators is impacting the availability and pricing of consumer-grade graphics cards.

Looking Ahead

As the AI landscape continues to evolve, the demand for both professional and consumer-grade hardware for AI applications will likely remain strong. The challenge for NVIDIA and other manufacturers will be to balance the needs of different market segments while optimizing production and supply chain management. Gamers, in the meantime, may need to brace themselves for continued price volatility and limited availability of high-end graphics cards.

References: