Specially developed for the unique requirements of AI
The NVIDIA DGX SuperPOD is far more than just a data centre – it is the technological blueprint for the modern ‘AI Factory’. At a time when generative AI models are reaching trillions of parameters, the SuperPOD delivers the necessary exascale performance in a fully integrated, ready-to-use environment.
Strategic advantage: Blackwell today, Rubin tomorrow
The AI landscape is evolving rapidly. With sysGen, you can rely on an infrastructure that guarantees top-tier performance today whilst also being future-proof:
- Blackwell generation (available now): Based on the GB200 NVL72, this system delivers real-time inference for LLMs that is up to 30 times faster than the H100 series.
- Rubin Platform (Expected in the second half of 2026): The next stage for ‘Agentic AI’. With the new Vera CPU (88 cores) and HBM4 memory, Rubin reduces the cost per inference token by a factor of 10.
Um sich an dieser Stelle das eingebettete YouTube Video anzeigen zu lassen, müssen Sie der Nutzung von "Externen Medien" zustimmen.
Alternativ können Sie das Video auch hier direkt bei Youtube in einem neuen Fenster öffnen.
Ideal choice for large development teams working on AI workloads within enterprises.
DGX SuperPOD with DGX GB200 systems is liquid-cooled, rack-scale AI infrastructure with predictive management capabilities purpose-built for training and inferencing trillion parameter generative AI models. With 72 NVIDIA Blackwell GPUs per rack, and scalability to tens of thousands of NVIDIA GB200 Grace Blackwell Superchips in total, it features full-stack resilience with an intelligent control plane, integrated hot spares and automated checkpointing and restart to maximize utilization.
The NVIDIA DGX SuperPOD with DGX Rubin NVL8 systems comprises 64 DGX Rubin NVL8 systems, each featuring 512 Rubin GPUs. NVIDIA DGX Rubin NVL8 systems bring the performance of Rubin with x86 CPUs into a liquid-cooled form factor, enabling businesses to efficiently enter the Rubin era for any AI project in the development and deployment pipeline. Each DGX Rubin NVL8 is powered by eight NVIDIA Rubin GPUs and sixth-generation NVLink, delivering 5.5 times the NVFP4 FLOPS compared to NVIDIA Blackwell systems.
Key features: Why the DGX SuperPOD makes all the difference
To take AI projects from development to production-scale deployment, a perfectly optimised infrastructure is required. The DGX SuperPOD offers key advantages in this regard:
Full-stack AI infrastructure
A ready-to-use supercomputer with an optimised design. Computing power, high-speed networking (Quantum-X800), data storage and software are seamlessly integrated as standard.
Massive supercomputing
Scalability to tens of thousands of GPUs enables the training and inference of state-of-the-art generative models without technological bottlenecks.
Maximum developer productivity
Thanks to company-wide cluster management (NVIDIA Base Command™) and optimised libraries, your experts can focus on model development rather than managing the infrastructure.
Resilience put to the test
Each system is pre-validated using real AI workloads. The intelligent control layer ensures full-stack resilience and consistent availability.
Exclusive expertise
As an NVIDIA Elite Partner, sysGen provides support throughout the entire lifecycle – from planning liquid cooling systems to performance optimisation.
Technical comparison
Merkmal | DGX SuperPOD (Blackwell) | DGX SuperPOD (Rubin) |
|---|---|---|
| GPU-Architektur | Blackwell | Rubin |
| Prozessor (CPU) | Grace CPU | Vera CPU (88 Kerne) |
| Speichertechnologie | HBM3e | HBM4 |
| Leistung (FP4) | ~1,4 ExaFLOPS pro Rack | 28,8 ExaFLOPS (Cluster) |
| Kühlung | Flüssigkeitsgekühlt | Flüssigkeitsgekühlt |
| Verfügbarkeit | Sofort verfügbar | Ab 2. Jahreshälfte 2026 |
Get in touch now!
Rely on the best AI infrastructure on the market
With the NVIDIA DGX SuperPOD, you get the technology that the world's leading companies and research institutions are already using successfully. Let's work together to realise your AI vision and prepare your company for the future.
Contact sysGen today and find out how the NVIDIA DGX SuperPOD can revolutionise your AI strategy. Together we can put your AI projects in the fast lane!
FAQ: Frequently asked questions about Blackwell and Rubin
What is the biggest advantage of a turnkey solution?
Time savings and risk minimisation. Instead of spending months integrating the network, storage and cooling systems, the SuperPOD is up and running within a few weeks and has been pre-validated for maximum performance.
- What is the ‘Vera CPU’?
Vera is NVIDIA’s latest data centre CPU. It has been specifically designed to eliminate the bottleneck in data transfer between the CPU and GPU, which is particularly crucial for ‘Agentic AI’ (AI agents that plan tasks autonomously).
- What is the advantage of the new Vera CPU in the Rubin series?
Vera was developed specifically for agent-based AI. It eliminates data bottlenecks between the CPU and GPU, thereby enabling more complex logical reasoning whilst significantly reducing energy consumption.
- Does my data centre need to be upgraded for Blackwell or Rubin?
Yes. Both systems use state-of-the-art liquid cooling (direct-to-chip). sysGen can assist you with planning the necessary cooling distribution units (CDUs) and the power supply.
- Why do both systems require liquid cooling?
The enormous density (72 GPUs per rack in Blackwell, up to 144 in Rubin) generates heat that cannot be efficiently dissipated using air cooling. Liquid cooling enables higher computing power in a smaller space and reduces operating costs (TCO).
- Is it worth waiting for Rubin, or should I invest in Blackwell now?
Blackwell is the ideal solution for businesses that need to scale productive AI workflows today. As the software (Mission Control) is compatible with both generations, investments in Blackwell can be seamlessly integrated into Rubin environments at a later date.
- Can I start with Blackwell and upgrade to Ruby later?
Absolutely. The NVIDIA architecture is designed for continuity. The management software is compatible across generations, which safeguards your investment in the long term.