AI for Edge and Departmental Workloads

Designed for smaller AI teams, edge environments, and pilot deployments, the entry-level iQ systems offer modular flexibility with up to 2 GPUs, NVMe-powered performance, and seamless integration with Hammerspace for data orchestration across environments.

Optionally, VSP One Object can be added for scalable, erasure-coded object storage—ideal for cost-efficient data retention.

Typical Use Cases:

• AI model prototyping
• Edge inference for video analytics or smart manufacturing
• Departmental AI/ML applications
• Academic or research environments

Scalable AI for Mid-Size Enterprise & Production Teams

The Hitachi iQ M Series offers a modular, scalable, and cost-effective AI infrastructure solution tailored for organizations seeking to balance performance with budget considerations. Designed to support a wide range of AI workloads, from initial deployments to full-scale production, the M Series integrates accelerated compute servers with robust networking and flexible storage options, including high-performance file systems and object storage.

Optionally, VSP One Object can be added for scalable, erasure-coded object storage—ideal for cost-efficient data retention.

Typical Use Cases:

• LLM training and fine-tuning
• Enterprise RAG pipelines and document AI
• Real-time AI inferencing
• AI-powered recommendation engines
• Image and video processing
• Industrial predictive maintenance
• Smart video analytics
• Enterprise document processing

Extreme Performance for Enterprise-Scale AI

Built for the most demanding AI workloads, the Hitachi iQ High-End systems combine NVIDIA DGX and HGX platforms with industry-leading throughput, low-latency networking, and full-stack AI infrastructure. Ideal for LLM training, GenAI production, delivering unmatched performance, resilience, and scalability.

By combining the extreme compute power of HGX or DGX systems with Hitachi Content Software for File (HCSF)—a high-performance parallel file system optimized for AI and GPU workloads—the iQ platform delivers the ultimate high-performance compute and data solution for enterprise AI.

Optionally, VSP One Object can be added for scalable, erasure-coded object storage—ideal for cost-efficient data retention.

Typical Use Cases:

• Foundation model training (LLM, GenAI, Vision AI)
• Enterprise RAG pipelines and document AI
• Real-time inferencing at scale
• Enterprise AI applications
• Image and video processing
• High-performance computing (HPC)

Unified Global Data Fabric for AI Anywhere

Hammerspace is a global file system that lets your AI workloads access data wherever it resides—on-premises, in other data centers, or in the cloud—without needing to move or copy that data manually. It presents all your storage as a single, unified namespace, regardless of where the physical data is stored.

Combined with Hitachi VSP One Block storage, it ensures that data is always available, resilient, and space-efficient—making it ideal for modern, distributed AI pipelines.

Key Benefits with Hitachi iQ:

• 100% data availability SLA
• 4:1 data reduction, no questions asked
• AI-ready data fabric across edge, core, and multi-cloud
• Real-time, zero-copy GPU access to remote datasets
• Global namespace with high-speed metadata acceleration
• Dedicated metadata engine for fast data placement and retrieval
• Automatic data orchestration & movement based on compute proximity
• Ideal for RAG pipelines, LLM tuning, and multi-tenant AI

Accelerated Storage for AI Workloads

Hitachi Content Software for File (HCSF) is a high-performance parallel, scale-out file system designed to deliver high throughput and low latency—ideal for AI and machine learning workloads. It is powered by the Weka® file system, providing the performance and flexibility needed to support demanding AI pipelines at enterprise scale.

In the Hitachi iQ platform, HCSF feeds large volumes of data into AI models quickly and efficiently. It supports POSIX-compliant file access, integrates with NVIDIA GPUDirect for direct storage-to-GPU communication, and avoids traditional I/O bottlenecks that slow down training or inferencing jobs.

When paired with Hitachi storage hardware and DGX or HGX systems, HCSF enables AI pipelines to operate at full speed—from data ingestion to model output.

Key Benefits:

• Ultra-low latency for AI/ML workloads
• Massive parallel throughput for large datasets
• Native POSIX file system with multiprotocol support
• Direct GPU access via NVIDIA GPUDirect Storage
• Scales linearly with additional nodes
• Optimized for AI training and inference