DDN, a reference in AI and data intelligence solutions, announced a major expansion of its AI & HPC data platform portfolio at ISC 2026, delivering breakthrough innovations across performance, efficiency, security, and cloud-scale AI infrastructure.The announcements include the launch of the new AI400X3M high-performance appliance, the official release of DDN’s distributed KV Cache acceleration technology integrated with Nvidia Dynamo, and new security, observability, and infrastructure efficiency enhancements for large-scale AI environments. 

As organizations race from AI pilots to production-scale AI operations, DDN is addressing the industry’s most critical bottlenecks across the entire AI data pipeline – from data ingestion and preparation to training, inference, RAG, and agentic AI. The company’s latest innovations are designed to maximize GPU utilization, accelerate inference performance, reduce infrastructure complexity, and improve AI economics by lowering cost per token and increasing tokens-per-watt efficiency across enterprise AI factories. 

“AI infrastructure is no longer just about compute. The economic success of AI depends on how efficiently organizations move, manage, secure, and operationalize data across the entire AI lifecycle,” said Alex Bouzari, CEO and co-founder, DDN. “At ISC 2026, DDN is introducing the next gen of AI data intelligence innovations designed to help customers maximize GPU utilization, reduce inference costs, accelerate time-to-token, and improve the overall economics of AI factories at massive scale.” 

Introducing the AI400X3M: Extreme Performance Density for AI and HPC 
Leading the announcements is the new DDN AI400X3M appliance, the latest evolution of DDN’s industry-leading EXAScaler platform. 

Designed for the most demanding AI and HPC environments, the AI400X3M delivers: 

• Up to 35% higher read throughput over the previous generation
• Up to 190GB /sec throughput performance to accelerate GPU access to data 
• Exceptional performance density in a compact footprint (up to 30 PB in a single rack) 
• Hybrid disk support for optimized economics and scalability, especially due to rising NAND flash costs 
• Extreme parallel throughput for supercomputing, training, inference, checkpointing, and large-scale AI pipelines 

The AI400X3M enables enterprises, sovereign AI programs, and cloud providers to dramatically increase infrastructure efficiency while reducing power, cooling, and operational costs. 

General availability is expected by the end of Q3 2026. 

Official Launch of DDN KV Cache Acceleration with Nvidia Dynamo Integration 
Following its preview at GTC 2026, DDN also announced the official launch of its distributed KV Cache acceleration architecture integrated with Nvidia Dynamo and available across DDN Infinia and EXAScaler AI data platforms. 

The solution dramatically accelerates large-scale AI inference by eliminating memory bottlenecks and enabling ultra-fast retrieval of model context directly from DDN’s AI-native data intelligence platform. 

Key capabilities include: 

• Shared distributed KV Cache fabric optimized for large-scale inference environments 
• Ultra-low latency data access for large-context inference and faster token generation 
• Optimized support for agentic AI, reasoning models, RAG, and multi-step inference pipelines 
• Deep integration with Nvidia Dynamo, vLLM, and modern inference frameworks 
• Improved GPU utilization and reduced idle compute cycles 
• Up to 55x faster KV cache loading performance for large-scale inference workloads 
• Lower cost per token and improved AI factory ROI through more efficient GPU and infrastructure utilization 

By moving KV cache closer to the data layer and reducing memory and networking bottlenecks, DDN enables enterprises and cloud providers to dramatically increase inference efficiency while reducing power consumption and infrastructure overhead associated with large-scale generative AI deployments. 

Accelerating the AI Data Pipeline from Training to Inference 
DDN’s latest innovations extend across the full AI data pipeline, helping enterprises operationalize AI faster and more efficiently from data preparation and model training to inference, RAG, reasoning, and agentic AI workflows. 

DDN Infinia delivers AI-native object storage engineered specifically for modern inference and retrieval-intensive workloads, providing ultra-low latency metadata performance, massive concurrency, and high-speed object access required for enterprise-scale AI factories. Combined with EXAScaler’s industry-leading parallel file system performance for training and checkpointing, DDN enables organizations to unify AI data infrastructure across the entire AI lifecycle. 

This architecture allows customers to eliminate data silos, maintain consistently high GPU utilization, accelerate time-to-first-token, and optimize AI infrastructure economics at scale. 

Additional Enterprise AI Infrastructure Enhancements 
DDN also introduced several platform enhancements focused on security, enterprise AI operations, multi-tenancy, and infrastructure observability, including: 

• Security 
• Bare-metal multi-tenancy  
• KMIP-based encryption and key management  
• VictoriaLogs integration for operational visibility  
• Multi-tenant APIs with and without CSI  
• Efficiency 
• Intelligent file pinning capabilities  
• NAND-accelerated Hot Pools to tier data from expensive all-flash drives to lower-cost HDDs  

The updates are designed to help enterprise, sovereign, and cloud AI operators improve workload isolation, governance, visibility, and infrastructure efficiency across production AI environments. 

Expanding Cloud AI Momentum with Managed Lustre and Salesforce 
DDN also highlighted continued momentum in cloud AI infrastructure, including new Managed Lustre innovations announced alongside Google Cloud Next and a new Salesforce deployment showcasing enterprise-scale AI performance and operational efficiency. 

The announcements further validate DDN’s leadership in enabling AI-native cloud architectures optimized for large-scale enterprise inference, RAG, and AI training workloads. 

DDN’s work helping Salesforce clear data bottlenecks offers a clear example of that momentum in action. With Google Cloud Managed Lustre, powered by DDN EXAScaler, Salesforce achieved 1.5x faster model training, a 75% reduction in I/O latency, and a 42% reduction in training costs. The results demonstrate how DDN is helping enterprise customers remove data bottlenecks and unlock greater productivity from every GPU, reducing the cost of AI while accelerating time to insight. As AI moves from experimentation to enterprise-scale deployment, DDN is providing the data intelligence foundation organizations need to maximize performance, efficiency, and return on AI investment.

Powering the World’s Largest AI Factories
DDN’s AI data intelligence platform powers many of the world’s largest and most advanced AI environments, including deployments supporting hyperscalers, sovereign AI initiatives, cloud providers, research institutions, and enterprise AI factories operating at massive GPU scale. 

By combining ultra-high-performance data infrastructure, AI-native orchestration, inference acceleration, observability, and operational efficiency, DDN continues to define the future of AI data intelligence, enabling organizations to build AI factories that maximize GPU ROI, reduce cost per token, and accelerate business outcomes across the full AI lifecycle.