DDN has announced its new Sovereign AI Blueprints and NVIDIA Reference Designs at the Supercomputing Conference 2025. This portfolio consists of validated, production-ready architectures intended for governments, enterprises, and research institutions to deploy artificial intelligence (AI) at national or industrial scale, with an emphasis on controlled performance, verified regulatory compliance, and energy efficiency. These solutions align with the NVIDIA AI Data Platform reference design and NVIDIA Certified Systems for AI factories, utilizing DDN’s unified data-intelligence software stack.
The Sovereign AI Blueprints and reference designs are built to ensure that data center and hyperscale operators can maintain data within defined jurisdictions, enforce local data governance policies, and leverage verifiable encryption and access controls. DDN claims these architectures deliver predictable high performance across training, inference, and retrieval augmented generation (RAG) workloads, while achieving over 99 percent sustained GPU utilization and up to 40 percent lower energy draw compared to legacy architectures.
Each blueprint provides a repeatable framework for deploying AI infrastructure at scale, offering features such as fine-grained multi-tenancy, per-tenant encryption, immutable storage, real-time telemetry, anomaly detection, and power-efficiency analytics. DDN highlights its AI400X3 AI Factory solutions as being engineered to maximize the performance of NVIDIA GB200 NVL72 and GB300 NVL72 systems, delivering up to 15 times faster checkpointing, four times higher ingest throughput, and sub-millisecond latency for demanding workloads. It also reports an 11 times improvement in performance per watt over legacy systems.
DDN notes that its solutions currently power large-scale AI projects like Yotta Shakti Cloud in India, which utilizes 8,000 NVIDIA Blackwell B200 GPUs on DDN architecture, reaching 99 percent GPU utilization and achieving 40 percent lower energy use. Other adopters include Singtel, which is leveraging DDN for next-generation AI services.
The Sovereign AI Blueprints and NVIDIA Reference Designs are offered globally via a network of partners, including system integrators (Accenture, Deloitte, Wipro, HCL Tech), cloud providers (CoreWeave, Nebius, Scaleway), and OEMs (Lenovo, Supermicro), supporting turnkey deployment and compliance frameworks for hyperscale and enterprise data centers.
“Sovereign AI is about owning your data destiny, not operating in silos,” said Sven Oehme, CTO at DDN. “DDN delivers the performance and efficiency organizations need to control their data, optimize GPU utilization, and scale AI sustainably.”
“AI-ready storage is no longer optional—it’s foundational to running at scale with data that moves at the speed of compute. Leveraging the NVIDIA AI Data Platform reference design, DDN powers AI factories with the performance, throughput, and scale needed to turn data into intelligence in real time,” said Justin Boitano, vice president, Enterprise AI Products, NVIDIA.
“By combining DDN’s data-intelligence software with Supermicro’s high-density NVIDIA systems, we enable fully validated, high-efficiency AI Factory and Sovereign AI deployments,” said Sachin Menon, Vice President, Cloud Engineering, Oracle.
Source: DDN
