VDURA (formerly Panasas)

VDURA, formerly Panasas, is a software-defined storage company that serves AI and high-performance computing (HPC) workloads.

Updated 6 Jun 2026 Reviewed by 7wData

Visit Website

Profile

Software-defined storage platform optimized for AI and high-performance computing clusters, using parallel file systems layered on object storage.

VDURA, formerly Panasas, is a software-defined storage company that serves AI and high-performance computing (HPC) workloads. Founded in 1999 by Garth Gibson and William Courtright, the company pioneered parallel file systems for network-attached storage. In May 2024, Panasas rebranded as VDURA—derived from "velocity" and "durability"—to reflect its strategic pivot from proprietary appliances to a software-centric subscription model.

The company has raised approximately $149 million across multiple funding rounds from investors including Mohr Davidow Ventures, The Carlyle Group, and Centennial Ventures. Ken Claffey became CEO in September 2023 and oversaw the complete reimagining of the legacy PanFS software stack into the VDURA Data Platform. In September 2025, co-founder Garth Gibson—a RAID pioneer who also led the Vector Institute—returned as Chief Technology and AI Officer to drive the company's vision for the AI era.

VDURA's primary offering is the VDURA Data Platform, which combines parallel file system performance with object storage resilience through a microservices architecture. The November 2025 release of Platform V12 introduced an Elastic Metadata Engine, snapshot support, and SMR HDD optimization, claiming 20% improvements in throughput and metadata performance. The company competes against established players like DDN, VAST Data, and WEKA in the AI-HPC storage market.

VDURA recently won a major U.S. defense GPU installation contract for one of the world's largest defense deployments, supporting 20 PB in phase one with planned expansion to 200 PB. The company claims to deliver 60% lower total cost of ownership than all-flash alternatives through tiered storage and commodity hardware flexibility. VDURA operates with approximately 49 to 107 employees (sources vary) across multiple continents and serves customers in life sciences, manufacturing, media and entertainment, energy, government, and academic research.

Who buys this

National laboratories and federal agencies running large-scale GPU deployments
Academic research institutions and supercomputing centers
Life sciences and genomics research organizations
Manufacturing and aerospace companies with intensive computational workloads
Media and entertainment studios requiring massive parallel storage for rendering and post-production

Publicly disclosed clients

U.S. Federal System Integrator (defense GPU installation)
Magseis Fairfield (geophysics and seismic data acquisition)
Stanford University
UC San Diego Center for Microbiome Innovation
Toyota (large language model workloads)
Valley Water (AI for water-flow forecasting)

Strengths and what to watch

Strengths

Founder return: Garth Gibson (RAID co-inventor, parallel file system pioneer) rejoined as CTAIO in September 2025, bringing foundational credibility to the AI infrastructure pivot.
Major defense win: Selected by a Tier-1 U.S. federal integrator for one of the world's largest defense GPU installations, validating product maturity and government-grade security requirements.
Cost positioning: Claims 60% lower TCO than all-flash competitors and launched Flash Relief Program in January 2026 to match VAST and WEKA quotes at half the total cost.

Watch for

Small team and customer concentration risk: Estimated 49–107 employees competing against much larger rivals (DDN, VAST, WEKA); limited publicly named customers outside defense and academic sectors suggests concentration on HPC niche.
Market maturity and benchmark gaps: Despite competitive claims, DDN dominates IO500 benchmarks, outperforming VDURA and rivals by up to 11x in real-world AI workloads; newer entrant status means limited independent customer references.
Organizational turbulence: Glassdoor reviews note poor HR and new management credibility issues; CEO transition from Faye Pairman (pre-2020) to Tom Shea (2020–2023) to Ken Claffey (2023-present) plus immediate Platform overhaul suggests execution risk despite founder return.

Recent moves

Key Information

Industry: Storage
Founded: 1999

Frequently Asked Questions

What is VDURA storage?

VDURA is a software-defined storage platform built for AI and HPC workloads. It combines parallel file system performance with object storage resilience through a microservices architecture, replacing Panasas's legacy appliance model with a flexible, subscription-based offering designed for modern compute clusters.

How much does VDURA cost compared to alternatives?

VDURA claims 60% lower total cost of ownership than all-flash alternatives through tiered storage and commodity hardware flexibility. The company launched its Flash Relief Program in January 2026 to match VAST and WEKA quotes at half the total cost, targeting price-sensitive HPC deployments.

Why did Panasas rebrand to VDURA?

Panasas rebranded to VDURA in May 2024 to signal its strategic pivot from proprietary hardware appliances to a software-centric, subscription-based model. The name derives from velocity and durability, reflecting the company's complete repositioning for modern AI and HPC infrastructure demands, abandoning legacy appliance constraints.

Who are VDURA's customers?

VDURA serves national laboratories, federal agencies, academic research institutions, life sciences organizations, manufacturing companies, and media studios. Notable customers include Stanford University, UC San Diego, Toyota, and a major U.S. federal system integrator for defense GPU deployments supporting 20+ petabytes of data.

What's new in VDURA Data Platform V12?

Platform V12, released in November 2025, introduced an Elastic Metadata Engine, snapshot support, and SMR HDD optimization. VDURA claims 20% improvements in throughput and 20x metadata acceleration. The platform achieves general availability in Q2 2026, targeting cost-conscious organizations building large-scale AI and HPC infrastructure.

How does VDURA compare to DDN, VAST, and WEKA?

VDURA competes on cost and software flexibility but faces benchmark challenges. DDN dominates IO500 benchmarks, outperforming VDURA by up to 11x in real-world AI workloads. VDURA's strengths lie in affordability and the recent return of founder Garth Gibson as CTO, bringing credibility but limited independent customer references.

How VDURA (formerly Panasas) compares

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

Positioning: Software-defined storage platform optimized for AI and high-performance computing clusters, using parallel file systems layered on object storage.
Customer segments: National laboratories and federal agencies running large-scale GPU deployments
Strengths: Founder return: Garth Gibson (RAID co-inventor, parallel file system pioneer) rejoined as CTAIO in September 2025, bringing foundational credibility to the AI infrastructure pivot.
Watch for: Small team and customer concentration risk: Estimated 49–107 employees competing against much larger rivals (DDN, VAST, WEKA); limited publicly named customers outside defense and academic sectors suggests concentration on HPC niche.
Recent moves: VDURA Unveils Data Platform V12 with Elastic Metadata Engine, Snapshot Support, and SMR HDD Optimization

Positioning: Dominant HPC and AI storage vendor, leads IO500 benchmarks, expanding from national labs into enterprise AI.
Customer segments: National labs, supercomputing centers, hyperscalers, neoclouds, enterprise AI teams; infrastructure architects managing GPU-dense deployments.
Strengths: Tops IO500 HPC storage rankings, with real-world AI workload benchmarks outpacing rivals by up to 11x.
Watch for: DDN acknowledges an enterprise sales gap as it expands beyond HPC research customers; enterprise go-to-market coverage is still being built.
Recent moves: January 2025: Blackstone invested $300 million at a $5 billion valuation to fund enterprise AI expansion.

Positioning: Unified all-flash platform collapsing storage, database, and compute layers, marketed as infrastructure OS for large-scale AI.
Customer segments: Hyperscalers, neoclouds, enterprise AI builders, large GPU cluster operators; infrastructure VPs and CTOs consolidating AI data tiers.
Strengths: Single shared-everything architecture integrating storage, database, and compute, sustaining over $500 million CARR with positive operating margins.
Watch for: All-flash architecture exposed to SSD price surge: a 25 PB deployment cost rose 189% between Q2 2025 and Q1 2026.
Recent moves: April 2026: $1 billion Series F closed at $30 billion valuation with NVIDIA as a participating investor.

Positioning: Flash-speed distributed file system for GPU-dense AI and HPC clusters, marketed on low-latency inference and GPU memory extension.
Customer segments: Fortune 50 enterprises, neoclouds, AI training infrastructure teams, GPU cluster operators; storage architects and AI infrastructure leads.
Strengths: Distributed NVMe file system positioned as GPU memory extension, adopted by 30% of Fortune 50 with NVIDIA backing.
Watch for: Valued at $1.6 billion versus VAST at $30 billion and DDN at $5 billion, limiting enterprise sales depth and resources.
Recent moves: June 2025: Launched NeuralMesh, an adaptive storage architecture purpose-built for agentic AI and exascale deployments.

Sources

www.vdura.com — Company overview, products (V12 Data Platform), market position, customer segments
blocksandfiles.com — Garth Gibson appointment as CTO, his background in RAID and parallel file systems, and strategic significance
www.businesswire.com — Rebranding from Panasas to VDURA in May 2024, software-defined subscription model transition
www.cbinsights.com — Funding history, total funding of $149M+ across multiple Series rounds, founding year 1999
insidehpc.com — Defense GPU installation contract, technical specifications (20 PB phase 1, 200 PB phase 2, 2.5 TBps)
blocksandfiles.com — Competitive positioning versus DDN, VAST Data, WEKA; market strategy and cost claims
www.vdura.com — V12 platform features, performance claims (20% throughput improvement, 20x metadata acceleration), Q2 2026 GA date
www.glassdoor.com — Employee feedback, HR and management concerns