# Hammerspace

> Hammerspace is the Data Platform for AI Anywhere: it unifies, orchestrates, and accelerates unstructured data across on-premises, hybrid, and multi-cloud environments through a single global namespace — without copying or migrating files between silos. Its core focus is making enterprises' existing, fragmented data "AI-ready" so GPU clusters, training pipelines, and inferencing/agentic workloads can use it immediately.

Hammerspace is not a storage array or a single cloud service — it's a metadata-driven software layer (the AI Data Platform, or "AIDP") that sits above existing storage (NAS, object, parallel file systems, cloud) and makes it behave as one system. Notes for interpreting the rest of this file:

- "AI-readiness" is Hammerspace's central pitch: fragmented, siloed data is framed as the main barrier to enterprise AI, and Hammerspace's approach is to "unify, not centralize" — cataloging data in place via Data-in-Place Assimilation rather than copying it into a new silo first.
- "Global namespace" is the core architectural concept: one logical, location-independent view of files and objects across sites, clouds, and storage vendors.
- "Tier 0" is Hammerspace's use of server-local NVMe on GPU compute nodes as a shared, orchestrated high-performance storage tier, aimed at maximizing GPU utilization and eliminating idle time (published results: 4.4x higher throughput, 56% lower latency, 66% faster checkpointing).
- Hammerspace uses standard protocols (NFS, SMB, S3) and open standards rather than proprietary clients/agents, so it integrates with existing GPU clusters and applications without rewriting them. Deployable as software, via cloud marketplace (AWS, Azure, GCP, OCI), or as a certified appliance (Dell, Supermicro, Hitachi Vantara).
- Primary audiences: enterprise IT/infrastructure teams, AI/ML platform teams, and HPC teams evaluating data orchestration, GPU utilization, and AI-readiness strategy.

## AI Data Platform (AIDP) & AI Readiness

- [AI Data Platform Overview](https://hammerspace.com/aidp-overview/): What AIDP is and how it makes existing data AI-ready without migration
- [AI Training](https://hammerspace.com/ai-training/): Feeding GPU clusters for large-scale model training
- [Sovereign AI](https://hammerspace.com/data-sovereignty/): Data residency and governance for regulated or sovereign AI deployments
- [Get AI-Ready by Putting Data First](https://hammerspace.com/get-ai-ready-by-putting-data-first/): The case for data readiness as the prerequisite to enterprise AI success
- [IDC Special Study: AI-Ready Data Storage Infrastructure](https://hammerspace.com/idc-special-study-on-ai-ready-data-storage-infrastructure/): Third-party research on what AI-ready infrastructure requires

## Platform Architecture

- [Platform Overview](https://hammerspace.com/platform-overview/): How the data platform, global namespace, and orchestration engine work together
- [Product Architecture](https://hammerspace.com/product-architecture/): Parallel file system architecture separating metadata from data
- [Global Namespace](https://hammerspace.com/global-namespace/): Unified, location-independent access across sites, clouds, and storage
- [Data-in-Place Assimilation](https://hammerspace.com/data-assimilation/): Instantly catalog and access existing data without moving it first
- [Data Orchestration](https://hammerspace.com/data-orchestration/): Policy-based automated data placement and movement
- [Tier 0](https://hammerspace.com/tier-0/): Server-local NVMe as a shared high-performance storage tier for GPU clusters
- [Competitive Advantage](https://hammerspace.com/hammerspace-advantage/): How Hammerspace differs from traditional storage and file system vendors

## Solutions

- [Artificial Intelligence](https://hammerspace.com/artificial-intelligence/): Data orchestration for AI/ML training, inferencing, and RAG pipelines
- [High-Performance Computing](https://hammerspace.com/high-performance-computing/): Data management for HPC clusters and research workloads
- [Hybrid Cloud](https://hammerspace.com/hybrid-cloud/): Moving compute and data seamlessly across data centers and AWS, Azure, GCP, and OCI
- [Life Sciences & Healthcare](https://hammerspace.com/life-sciences/): Genomics, imaging, and research data workflows
- [Media & Entertainment](https://hammerspace.com/media-entertainment/): VFX, animation, and render pipeline data management
- [Financial Services](https://hammerspace.com/financial-services/): Low-latency, compliant data infrastructure for trading and analytics
- [Public Sector](https://hammerspace.com/public-sector/): Federal and government data orchestration and sovereignty

## Proof Points

- [IO500 Benchmark Results](https://hammerspace.com/io500-benchmark-results/): Independently validated results showing 2x HPC-class AI performance with under half the infrastructure
- [Customers](https://hammerspace.com/customers/): Case studies including Netflix, AMD, Meta, PayPal, Los Alamos National Laboratory, and Vanderbilt ACCRE
- [Awards and Recognition](https://hammerspace.com/awards/): Industry awards and analyst recognition

## Company

- [About Us](https://hammerspace.com/about-us/): Company background and leadership, including the original architects of NVMe and Linux NFS
- [Newsroom](https://hammerspace.com/news-room/): Press releases and company news
- [Careers](https://hammerspace.com/careers/): Open roles and company culture
- [Contact Us](https://hammerspace.com/contact-us/): Sales inquiries and live demo requests

## Optional

- [Blog](https://hammerspace.com/blog/): Technical articles on AI data orchestration, Tier 0, and hybrid cloud architecture
- [Resource Library](https://hammerspace.com/resources/): White papers, solution briefs, case studies, and technical documentation
- [Data Unchained Podcast](https://hammerspace.com/data-unchained-podcast/): Interviews on AI data infrastructure topics
- [Privacy Statement (US)](https://hammerspace.com/privacy-statement-us/): Data privacy practices
- [Support](https://hammerspace.com/support/): Customer support