SuperX Launches Rack-Scale AI Platform Powered by NVIDIA GB300 Chips, Redefining Data Center Infrastructure

Rhea-AI Summary

SuperX (NASDAQ: SUPX) launched the SuperX GB300 NVL72 System on October 16, 2025, a liquid-cooled, rack-scale AI platform built around the NVIDIA GB300 Grace Blackwell Ultra Superchip.

The NVL72 rack links 72 Blackwell Ultra GPUs and 36 Grace CPUs to deliver up to ≈1.8 exaFLOPS FP4 AI compute, ≈165TB HBM3E and ≈17TB LPDDR5X memory per rack, 800Gb/s InfiniBand XDR networking, and a 48U MGX rack form factor. SuperX positions the system for hyperscale, sovereign AI, exascale scientific computing and industrial digital twins, and bundles the rack with liquid cooling and 800VDC power infrastructure as part of a prefabricated modular AI factory solution.

Positive

• Up to ≈1.8 exaFLOPS FP4 AI compute per rack
• ≈165TB HBM3E total GPU memory per rack
• Built-in liquid cooling for continuous 24/7 operation
• 800Gb/s InfiniBand XDR networking for low latency

Negative

• Requires 800VDC power infrastructure for deployment
• High compute density forces legacy air-cooled data center upgrades
• Prefabricated solution increases deployment capex and complexity

Insights

Data Center Infrastructure Analyst

SuperX introduced a liquid-cooled, rack-scale GB300 NVL72 system delivering up to 1.8 exaFLOPS per rack, shifting infrastructure needs toward high-voltage DC and liquid cooling.

The system pairs 72 NVIDIA Blackwell Ultra GPUs with 36 NVIDIA Grace CPUs to concentrate compute and memory (≈165TB HBM3E, ≈17TB LPDDR5X) into a 48U rack. This design trades traditional air-cooled, AC-centric assumptions for liquid cooling and 800VDC power distribution to support sustained, high-density AI workloads. The architecture emphasizes unified memory and a 900GB/s chip link, which address common I/O bottlenecks when training very large models.

Primary dependencies include data center power delivery and cooling upgrades; legacy AC power and air cooling are described as inadequate for continuous operation at this density. Risks center on site readiness and capital expenditure to deploy 800VDC and liquid cooling at scale, plus integration of the prefabricated modular AI factory stack. Short-term checkpoints to watch include customer adoption announcements and proof-of-deployment timelines; expect tangible deployment signals within the next 6–18 months from Oct. 16, 2025.

SINGAPORE, Oct. 16, 2025 /PRNewswire/ -- SuperX AI Technology Limited (NASDAQ: SUPX) ("the Company" or "SuperX") today announced the launch of the SuperX GB300 NVL72 System, a groundbreaking, rack-scale AI supercomputing platform powered by the NVIDIA GB300 Grace Blackwell Ultra Superchip. Designed to conquer the physical and computational limits of training and deploying next-generation trillion-parameter models, the liquid-cooled GB300 System delivers a significant leap in performance density and energy efficiency that redefines the requirements for modern data center infrastructure.

The arrival of rack-scale AI systems like the SuperX GB300 NVL72 system marks a critical inflection point for the industry. By delivering up to 1.8 exaFLOPS of AI performance in a single, liquid-cooled rack, the system achieves a level of compute density that the traditional air-cooled data center designs and conventional Alternate Current (AC) power distribution find challenging to support. This immense concentration of compute and power in a smaller footprint means that legacy infrastructure cannot adequately support these next-generation workloads.

This technological shift elevates advanced power solutions, particularly 800 Voltage Direct Current (800VDC), from an efficiency advantage to a fundamental necessity. The ability to deliver massive amounts of power directly and efficiently to the rack is now critical for stability, safety, and operational viability. The SuperX GB300 NVL72 System is therefore positioned not as a standalone product, but as the centerpiece of our full-stack SuperX Prefabricated Modular AI Factory solution that includes the essential liquid cooling system, and 800VDC power infrastructure required for its deployment.

The Grace Blackwell Ultra Superchip Advantage

The SuperX GB300 System is built upon the foundational GB300 Superchips with 72 NVIDIA Blackwell Ultra GPUs and 36 NVIDIA Grace CPUs in a 2-to-1 combination. This integration delivers extreme bandwidth through a 900GB/s chip-to-chip link that seamlessly connects the Grace CPU's high-memory bandwidth to the Blackwell Ultra GPUs. It ensures unified memory by combining 2,304GB of HBM3E with the Grace CPU's expansive LPDDR5X, enabling the handling of the largest models and K/V caches without I/O bottlenecks. At the same time, it achieves optimal efficiency, as the Grace CPU's power-efficient processing complements Blackwell Ultra GPU's compute-intensive performance, providing superior performance per watt.

Rack-Scale Exascale AI

The GB300 System is natively designed for massive horizontal scaling. A single SuperX GB300 System can scale up to the NVL72 rack configuration, linking 72 Blackwell Ultra GPUs together into a single, massive GPU system. Delivering breakthrough performance, the system achieves up to 1.8 exaFLOPS of FP4 AI compute within a single rack, redefining industry standards for large-scale AI training and inference. With 800Gb/s InfiniBand XDR connectivity, it ensures ultra-low latency across the most demanding AI clusters, unlocking significant scalability potential. To maintain this level of performance, the system incorporates an advanced liquid cooling design, enabling exceptional density and continuous 24/7 operation while maximizing energy efficiency.

Technical Specifications:

CPU (Per Rack)	36* NVIDIA Grace CPUs (144 Arm Neoverse V2 cores total)
GPU (Per Rack)	72* NVIDIA Blackwell Ultra GPUs
Total HBM3E Memory	≈165TB (GPU High-Bandwidth Memory)
Total LPDDR5X Memory	≈17TB (Grace CPU Memory)
Peak AI Performance	≈1.8 ExaFLOPS (FP4 AI)
Networking	4* NVIDIA NVLink Connectors (1.8TB/s) 4* NVIDIA ConnectX 8 OSFP Ports (800Gb/s) 1* NVIDIA BlueField 3 DPUs (400Gb/s)
Dimension	48U NVIDIA MGX Rack 2296mm(H) x 600mm(W) x 1200mm(D)

Market Positioning

The SuperX GB300 NVL72 System is the ideal infrastructure for organizations building the foundation of tomorrow's AI:

• Hyperscale & Sovereign AI: For constructing national AI infrastructure, public cloud services, and massive enterprise AI factories that require exascale compute to train and serve the most complex multi-modal and large language models.
• Exascale Scientific Computing: For governments and research institutions tackling grand challenges in physics, materials science, and climate modeling that necessitate highly efficient, rack-scale compute.
• Industrial Digital Twins: For automotive, manufacturing, and energy sectors building large-scale, high-fidelity digital twins requiring the combined processing power of the Grace CPU and Blackwell Ultra GPU.

Purchase & Contact Information

For product inquiries, sales, and detailed specifications, please contact our product sales team at: Sales@superx.sg

About SuperX AI Technology Limited (NASDAQ: SUPX)

SuperX AI Technology Limited is an AI infrastructure solutions provider, offering a comprehensive portfolio of proprietary hardware, advanced software, and end-to-end services for AI data centers. The Company's services include advanced solution design and planning, cost-effective infrastructure product integration, and end-to-end operations and maintenance. Its core products include high-performance AI servers, 800 Voltage Direct Current (800VDC) solutions, high-density liquid cooling solutions, as well as AI cloud and AI agents. Headquartered in Singapore, the Company serves institutional clients globally, including enterprises, research institutions, and cloud and edge computing deployments. For more information, please visit www.superx.sg

Safe Harbor Statement

This press release may contain forward-looking statements. In addition, from time to time, we or our representatives may make forward-looking statements orally or in writing. We base these forward-looking statements on our expectations and projections about future events, which we derive from the information currently available to us. You can identify forward-looking statements by those that are not historical in nature, particularly those that use terminology such as "may," "should," "expects," "anticipates," "contemplates," "estimates," "believes," "plans," "projected," "predicts," "potential," or "hopes" or the negative of these or similar terms. In evaluating these forward-looking statements, you should consider various factors, including: our ability to change the direction of the Company; our ability to keep pace with new technology and changing market needs; and the competitive environment of our business. These and other factors may cause our actual results to differ materially from any forward-looking statement.

Forward-looking statements are only predictions. The reader is cautioned not to rely on these forward-looking statements. The forward-looking events discussed in this press release and other statements made from time to time by us or our representatives, may not occur, and actual events and results may differ materially and are subject to risks, uncertainties, and assumptions about us. We are not obligated to publicly update or revise any forward-looking statement, whether as a result of uncertainties and assumptions, the forward-looking events discussed in this press release and other statements made from time to time by us or our representatives might not occur.

Contact Information
Product Inquiries: sales@superx.sg
Investor Relation: ir@superx.sg 

Follow our social media:
X.com: https://x.com/SUPERX_AI_ 
LinkedIn: https://www.linkedin.com/company/superx-ai
Facebook: https://www.facebook.com/people/Super-X-AI-Technology-Limited/61578918040072/# 

View original content to download multimedia:https://www.prnewswire.com/news-releases/superx-launches-rack-scale-ai-platform-powered-by-nvidia-gb300-chips-redefining-data-center-infrastructure-302586079.html

SOURCE SuperX AI Technology Ltd

FAQ

What performance does the SuperX GB300 NVL72 System (SUPX) deliver?

The system delivers up to ≈1.8 exaFLOPS FP4 AI compute in a single rack.

What hardware makes up the SUPX GB300 NVL72 rack?

Each rack includes 72 NVIDIA Blackwell Ultra GPUs and 36 NVIDIA Grace CPUs with ≈165TB HBM3E and ≈17TB LPDDR5X memory.

What infrastructure is required to deploy the SuperX GB300 NVL72 (SUPX)?

Deployment requires integrated liquid cooling and 800VDC power distribution to support the rack's power density.

Who is the target customer for the SuperX GB300 NVL72 System (SUPX)?

Target customers include hyperscalers, sovereign AI projects, governments, research institutions, and enterprises building large digital twins.

How does the SuperX GB300 NVL72 handle networking for large AI clusters?

It uses 800Gb/s InfiniBand XDR, NVIDIA NVLink, and BlueField DPUs to enable ultra-low latency and horizontal scaling.

How can investors or customers get sales information for SUPX GB300?

Product and sales inquiries are directed to the SuperX product sales team at Sales@superx.sg.