Startups > Hammerspace

Hammerspace

Hammerspace helps enterprises orchestrate unstructured data across storage systems to accelerate AI workloads.

Series B $156.7M total Founded 2018 San Mateo, California 183 employees

Hammerspace is a metadata-driven data orchestration platform that sits above file and object storage systems, providing a unified global namespace for accessing unstructured data across on-premises and public cloud infrastructure. Rather than storing data itself, it acts as a software layer that enables enterprises to access files regardless of storage type, vendor, or location while optimizing performance for AI and HPC workloads. The platform automatically transitions data between high-performance Tier 0 storage during compute-intensive operations and cost-efficient object storage during idle periods, eliminating GPU idle time caused by data bottlenecks.

Problem solved

GPU clusters sitting idle waiting for data to be ingested and processed, data scattered across incompatible storage systems with no unified access, inability to efficiently manage data movement between expensive high-performance and cost-effective archival storage.

Target customer

Enterprises with significant unstructured data across multiple storage systems, hyperscalers, HPC/AI organizations, government agencies, media and entertainment companies with large-scale GPU infrastructure.

Website LinkedIn Crunchbase Twitter / X

Founders

David Flynn

CEO and Co-founder

Previously CEO and co-founder of Fusion-io (acquired by SanDisk for $1.1B in 2014), specialized in high-performance computing and enterprise flash storage. Self-financed Hammerspace's early growth.

Trond Myklebust

CTO and Co-founder

20+ year maintainer and lead developer of the Linux kernel NFS client; continues to lead NFS protocol development.

Douglas Fallstrom

Co-founder

Limited public information available.

Funding history

Pre-Series A Unknown 2018-2023 Led by David Flynn + high net worth individuals · Long-term investing sources

Series A $56.7M July 2023 Led by Prosperity7 Ventures (Saudi Aramco) · Unknown

Series B $100M April 2025 Led by Altimeter Capital · ARK Invest (Cathie Wood's ARK Venture Fund)

Total raised: $156.7M

Industries

Artificial Intelligence (AI) Cloud Data Services Cloud Storage Data Management Data Storage Information Technology Software

Pricing

Subscription-based with tiered pricing based on data orchestration needs and usage. Enterprise licensing available for customized solutions with dedicated support. AWS Marketplace offering uses consumption-based pricing with no end date on subscriptions.

Notable customers

Blue Origin, National Science Foundation, Royal Caribbean Group, Meta Platforms, U.S. Department of Defense, Los Alamos National Laboratory, Hitachi Vantara, Supermicro, major media/entertainment companies (Star Wars, Stranger Things special effects production)

Integrations

AWS Marketplace, Linux kernel NFS protocol (maintained by CTO), Hitachi Vantara solutions, Supermicro infrastructure partnerships

Tech stack

GSAP (JavaScript frameworks) LazySizes (JavaScript libraries) jQuery Migrate (JavaScript libraries) jQuery (JavaScript libraries) YouTube (Video players) Webpack Module Federation RSS PWA Open Graph WordPress (Blogs) Slider Revolution (Widgets) Zoominfo (Analytics) Google Ads Conversion Tracking (Analytics) Microsoft Clarity (Analytics) Google Analytics (Analytics) HSTS (Security) Google Font API (Font scripts) Nginx (Web servers) Varnish (Caching) PHP (Programming languages) Microsoft 365 (Email) Fastly (CDN) Salesforce Marketing Cloud Account Engagement (Marketing automation) MySQL (Databases) MariaDB (Databases) Google Ads (Advertising) Google Tag Manager (Tag managers) Divi (Page builders) The SEO Framework (SEO) Pantheon (PaaS) Complianz (A/B Testing) New Relic (RUM) The Events Calendar (WordPress plugins) WP-PageNavi (WordPress plugins) EWWW Image Optimizer (WordPress plugins)

Website

hammerspace.com/

Competitors

WEKA

Focused parallel file system for HPC/AI; less emphasis on multi-cloud orchestration and metadata-driven global namespace.

Data Direct Networks (DDN)

Traditional enterprise storage vendor with broader hardware offerings; less specialized in metadata-driven orchestration.

VAST Data

Flash-based all-NVMe storage architecture; different approach to performance optimization versus Hammerspace's tiered orchestration.

Vdura

Competitor in AI data infrastructure; less information on specific architectural differences.

Why this matters: Hammerspace is backed by top-tier investors (Altimeter Capital, ARK Invest, Saudi Aramco) and addresses a critical bottleneck in AI infrastructure—the disconnect between expensive GPU horsepower and data availability. With a founding team combining HPC expertise (Flynn's Fusion-io exit) and Linux kernel NFS authority (Myklebust), the company is well-positioned to capture share in the rapidly expanding AI infrastructure market as enterprises move toward heterogeneous multi-cloud environments.

Best for: Enterprises running GPU-intensive AI/ML workloads or HPC simulations that need to maximize compute resource utilization by eliminating data access bottlenecks, and organizations managing unstructured data across heterogeneous storage environments.

Use cases

Accelerating AI Training

Machine learning teams use Hammerspace to feed massive training datasets into GPU clusters at Tier 0 performance speeds, eliminating idle GPU time. After training completes, data automatically migrates to cost-efficient object storage. This directly addresses the expensive GPU resource utilization problem.

Multi-Cloud Data Access

Enterprises with data spread across on-premises NAS, AWS S3, and Google Cloud Storage can present a unified namespace through Hammerspace, allowing applications to access any file seamlessly regardless of actual storage location. This eliminates the need for complex application-level logic to handle multiple storage backends.

Media/Entertainment Effects Processing

Production studios managing petabytes of video and effects data can orchestrate high-speed access to render farms for real-time processing, then automatically tier completed sequences to archive storage. Used in production of major films and streaming shows.

Government HPC Workloads

National labs running climate modeling, pandemic preparedness, and scientific computing can optimize data movement to/from supercomputers, ensuring continuous compute capacity utilization rather than waiting for data staging.

Alternatives

NetApp Broader enterprise storage vendor with file and object offerings; less specialized in AI/GPU optimization and metadata-driven orchestration.

Dell EMC Traditional enterprise storage and infrastructure company; different approach to data management versus specialized orchestration layer.

Pure Storage Flash-focused storage vendor; less emphasis on multi-vendor orchestration and global namespace abstractions.

FAQ

What does Hammerspace do? +

Hammerspace is a metadata-driven data orchestration platform that creates a unified global namespace above existing storage systems (on-premises and cloud). It enables seamless access to unstructured data regardless of storage location or type, and automatically optimizes data movement between high-performance and cost-efficient tiers based on usage patterns. It's particularly optimized for AI and HPC workloads to eliminate GPU idle time.

How much does Hammerspace cost? +

Hammerspace uses subscription-based pricing with tiers based on data orchestration needs and usage. Enterprise licensing with custom solutions and dedicated support is available. AWS Marketplace pricing is consumption-based with flexible subscription terms. Exact pricing requires contacting the company.

What are alternatives to Hammerspace? +

WEKA (parallel file system for HPC/AI), Data Direct Networks DDN (enterprise storage), VAST Data (all-NVMe storage), Vdura (AI data infrastructure), and traditional vendors like NetApp, Dell EMC, and Pure Storage offer overlapping capabilities with different architectural approaches.

Who uses Hammerspace? +

Customers include hyperscalers (Meta), government agencies (NSF, DoD, Los Alamos National Lab), media/entertainment companies (for visual effects production), aerospace companies (Blue Origin), cruise lines (Royal Caribbean), and enterprise infrastructure partners (Hitachi Vantara, Supermicro). Target customers are organizations with large unstructured data, GPU infrastructure, or complex multi-cloud environments.

How does Hammerspace compare to WEKA? +

WEKA is a parallel file system optimized for HPC/AI performance, while Hammerspace is a metadata-driven orchestration layer above existing storage. Hammerspace emphasizes unified access across heterogeneous storage vendors and automatic tiering between performance and cost-efficiency. WEKA focuses on ultra-high performance through a unified architecture but requires migration of data into its system.