Hammerspace

Updated

Founded in 2018 and based in San Mateo, California, Hammerspace is a standards-based data orchestration platform built to streamline AI and high-performance computing workloads.

Reviewed by 7wData

On this page

Profile

A unified data platform that gives AI and HPC workloads instant access to data across on-premises storage, public clouds, and edge environments without copying or migrating.

Founded in 2018 and based in San Mateo, California, Hammerspace is a standards-based data orchestration platform built to streamline AI and high-performance computing workloads. The company was started by industry veterans with deep infrastructure pedigree: David Flynn (who previously founded Fusion-io, the flash memory company acquired by SanDisk for $1.4B in 2014), Trond Myklebust (principal Linux NFS maintainer), and Douglas Fallstrom. This founding team, augmented by storage architects and Linux kernel contributors, gives Hammerspace unusual credibility in the open-source infrastructure community.

Hammerspace's core insight is that enterprises running AI training, inference, and high-performance computing are frustrated by data fragmentation across on-premises storage, public clouds, and edge environments. Their Global Data Environment provides a unified namespace spanning all three, allowing GPU clusters and compute nodes to access data at high performance without copying or migrating. The Tier 0 solution is their newest product: it transforms local NVMe storage in GPU server clusters into a shared, high-speed storage tier with up to 10x faster performance than traditional network storage.

Financially, the company has momentum. In 2024, Hammerspace achieved a tenfold increase in revenue and grew its customer base by 32 percent, fueled by adoption among hyperscalers and government agencies. Meta uses Hammerspace for Llama training infrastructure, while the Department of Defense, National Science Foundation, Blue Origin, and Royal Caribbean are also customers. The company raised $56.7M in Series A from Prosperity7 Ventures, ARK Invest, and others in 2023, followed by a $100M Series B in April 2025 led by Altimeter Capital and ARK Invest, valuing the company at over $500M.

CEO David Flynn has indicated an IPO target around 2027, though industry analysts are skeptical, suggesting acquisition may be a more likely exit path given the company's narrow technical focus and specific customer base. The company has been expanding globally, launching in Korea, Japan, and China in 2025. Recent strategic moves include contributions to the Linux kernel for NFSv4.2 protocol enhancements and a partnership with NVIDIA's AI Data Platform reference design. By February 2026, Hammerspace was recognized as "Innovation of the Year" by StorageNewsletter.

Track Hammerspace and 240+ vendors.

335k+ subscribers read the daily AI & data note. One email, both newsletters. Unsubscribe anytime.

Who buys this

  • Hyperscale AI training operations (Meta, major cloud providers)
  • U.S. government and defense agencies (DoD, NSF, research institutions)
  • Media and entertainment studios managing distributed visual effects workflows
  • Enterprise HPC environments requiring multi-site, multi-cloud data orchestration
  • Cloud service providers (Oracle, AWS, Azure, GCP) offering AI infrastructure

Publicly disclosed clients

  • Meta (Llama LLM training)
  • Department of Defense
  • National Science Foundation
  • Blue Origin
  • Royal Caribbean Group
  • Mathematic Studio

Strengths and what to watch

Strengths

  • Founding team's deep infrastructure credibility: CEO is Fusion-io founder; CTO is Linux NFS protocol maintainer; team includes Linux kernel contributors
  • Rapid commercial momentum: 10x revenue growth in 2024, 32% customer growth, achieved $500M+ valuation in Series B
  • Strategic partnerships with major platforms (NVIDIA AI Data Platform, Oracle Cloud Marketplace, Hitachi Vantara) expanding distribution and credibility

Watch for

  • IPO skepticism: Industry analysts (Omdia, Forrester) suggest acquisition is more likely than public markets exit given narrow technical focus
  • Customer concentration risk: Current customer base heavily weighted to hyperscale AI, government, and HPC; limited enterprise IT penetration
  • Competitive pressure from mature vendors (VAST Data, WEKA, NetApp, Pure Storage) with larger installed bases and broader product portfolios

Recent moves

Key Information

Industry
Data Orchestration
Founded
2018
Employees
201-500
Headquarters
San Mateo, CA

Frequently Asked Questions

What is Hammerspace?

Hammerspace is a unified data platform that gives AI and HPC workloads instant access to data across on-premises storage, public clouds, and edge environments without copying or migrating. Founded in 2018 by infrastructure veterans including Fusion-io founder David Flynn, it serves hyperscalers and government agencies.

Who founded Hammerspace?

Hammerspace was founded in 2018 by David Flynn, who previously founded Fusion-io (acquired by SanDisk for $1.4B in 2014), Trond Myklebust (principal Linux NFS protocol maintainer), and Douglas Fallstrom. The team also includes storage architects, Linux kernel contributors, and infrastructure engineers.

How does Hammerspace's Tier 0 work?

Hammerspace's Tier 0 transforms local NVMe storage in GPU server clusters into shared, high-speed storage accessible to all nodes. It delivers up to 10x faster performance than traditional network storage, 2.5x faster reads, 2x higher writes, and 51% lower latency compared to standard solutions.

Who uses Hammerspace?

Hammerspace serves hyperscale AI training operations, including Meta for Llama LLM training, as well as U.S. government agencies like the Department of Defense, National Science Foundation, and research institutions. Customers also include Blue Origin, Royal Caribbean, cloud providers, and media studios managing distributed workflows.

How much funding did Hammerspace raise?

Hammerspace raised $56.7M in Series A from Prosperity7 Ventures and ARK Invest in 2023. In April 2025, it secured $100M in Series B led by Altimeter Capital, achieving over $500M valuation. The company achieved 10x revenue growth in 2024 and grew its customer base 32 percent.

Is Hammerspace going public?

CEO David Flynn has indicated an IPO target around 2027, but industry analysts including Omdia and Forrester believe acquisition is more likely given Hammerspace's narrow technical focus and concentrated customer base in hyperscale AI, government, and HPC rather than enterprise IT.

How Hammerspace compares

Direct head-to-head against 3 competitors. Picked by 7wData.

This company

Hammerspace

Positioning
A unified data platform that gives AI and HPC workloads instant access to data across on-premises storage, public clouds, and edge environments without copying or migrating.
Customer segments
Hyperscale AI training operations (Meta, major cloud providers)
Strengths
Founding team's deep infrastructure credibility: CEO is Fusion-io founder; CTO is Linux NFS protocol maintainer; team includes Linux kernel contributors
Watch for
IPO skepticism: Industry analysts (Omdia, Forrester) suggest acquisition is more likely than public markets exit given narrow technical focus
Recent moves
Hammerspace posts nearly 14x YTD bookings over full-year 2025 as AI shifts to inference and physical AI

VAST Data

Positioning
AI data platform combining GPU-accelerated storage, analytics, and inference in a unified architecture for enterprise and cloud scale.
Customer segments
Large enterprises, national labs, cloud providers running GPU AI/ML or HPC. Buyer: VP or Director of Infrastructure.
Strengths
DASE architecture runs CUDA-accelerated storage natively on NVIDIA GPU fabric with multiprotocol (file, object, block) support and no disaggregation penalty.
Watch for
NetApp IP lawsuit (September 2025) against acquired Red Stapler executives is a live governance and vendor-stability risk for procurement committees.
Recent moves
April 2026 Series F closed at $30B valuation with NVIDIA participation, raising cumulative contracted ARR past $500M and total bookings past $4B.

WEKA

Positioning
Software-defined parallel file system eliminating GPU storage bottlenecks in AI training, inference, and HPC workloads without proprietary hardware.
Customer segments
Enterprise AI labs, neoclouds, GPU-cluster operators, national labs. Buyer: VP or Director of AI Infrastructure at GPU-dense organizations.
Strengths
NVMe-based distributed file system runs on any hardware, delivering sub-millisecond latency directly to GPU memory with no hardware vendor lock-in.
Watch for
Three C-suite exits (President, CFO twice, CRO) in under six months through mid-2025, signaling governance and leadership continuity risk.
Recent moves
June 2025: launched NeuralMesh, a re-architected adaptive mesh storage platform targeting agentic AI and AI factory deployments.

DDN

Positioning
IO500 and MLPerf benchmark leader for AI and HPC storage targeting supercomputing centers and enterprise AI infrastructure teams.
Customer segments
HPC architects and data center engineers at national labs, hyperscalers, and enterprise IT teams building production AI platforms.
Strengths
EXAScaler commercial Lustre parallel file system with decades of kernel-level tuning for NVMe and GPU-direct I/O at supercomputer scale.
Watch for
CMO departed May 2026 weeks after a new President/CRO was installed, two senior GTM roles turning over simultaneously.
Recent moves
January 2025: Blackstone invested $300M at a $5B valuation, DDN's first outside funding in 25 years, targeting healthcare, autonomous vehicles, and enterprise AI.

Sources

  1. techcrunch.com — Series B funding round ($100M), valuation ($500M+), customer names (Meta, DoD), growth metrics, IPO timeline
  2. www.techtarget.com — Series B details, analyst skepticism on IPO vs acquisition, competitive positioning, market focus
  3. hammerspace.com — Company mission, product portfolio (Global Data Environment, Tier 0, AI Data Platform), founding team pedigree, deployment options, use cases
  4. blocksandfiles.com — Series B funding details, CEO David Flynn background, competitive differentiation, Global Data Environment positioning
  5. hammerspace.com — Oracle partnership, Tier 0 performance metrics (2.5x faster reads, 2x higher writes, 51% lower latency), OCI deployment
  6. hammerspace.com — Geographic expansion (Korea, Japan, China launches in 2025), 32% YoY customer growth, 10x revenue growth metrics
  7. insidehpc.com — Korea market entry date (July 8, 2025), APAC expansion strategy, company growth metrics
  8. www.storagenewsletter.com — StorageNewsletter Innovation of the Year award (February 2026), market disruption claims, data platform capabilities