What is VAST Data?
The AI Operating System for data infrastructure at exabyte scale
- Category
- AI Data Infrastructure
- Headquarters
- New York, NY (R&D: Tel Aviv, Israel)
- Founded
- 2016
- Employees
- ~1,150 (2025)
- Total Funding
- ~$1.4B across 7+ rounds
- Valuation
- $30B (April 2026 Series F)
What is VAST Data?
VAST Data is the company behind the VAST AI Operating System — a unified software platform that collapses storage, database, distributed compute, and agentic AI orchestration into a single shared-everything fabric purpose-built for AI workloads. Founded in January 2016 and headquartered in New York City with a major R&D center in Tel Aviv, VAST has grown from a flash-storage disruptor into one of the fastest-scaling enterprise infrastructure companies in history, reaching approximately $2 billion in ARR run-rate by end of 2025 and more than $500 million in committed ARR (CARR) entering fiscal 2026.
VAST's core architecture — Disaggregated Shared-Everything (DASE) — separates stateless front-end compute nodes (CNodes) from all-flash NVMe storage enclosures (DNodes) connected over NVMe-over-Fabrics (NVMe-oF). This eliminates storage tiers, data movement pipelines, and traditional metadata bottlenecks, delivering sub-millisecond latency at petabyte-to-exabyte scale. The platform supports file, object, block, and database access simultaneously under a single namespace, speaking NFS v3/v4, SMB, S3, and Kubernetes CSI natively — making it the de facto storage layer for the world's most demanding GPU clusters.
Beyond storage, VAST has expanded into a full AI Operating System spanning six software domains: DataStore (all-flash distributed file and object storage), DataBase (vector and columnar database), DataEngine (distributed Python compute), DataSpace (global metadata and policy namespace), AgentEngine (multi-agent AI orchestration with NVIDIA integration), and SyncEngine (universal cloud replication to AWS, Azure, and GCP). At VAST Forward 2026 (February 25, 2026), the company added two new services — PolicyEngine (zero-trust governance for AI agents) and TuningEngine (structured learning loops for continuous model improvement) — completing a closed-loop AI operating environment.
Named customers include xAI's Colossus supercomputing cluster (200,000+ NVIDIA GPUs), CoreWeave (a $1.17 billion commercial partnership signed November 2025), Lambda, Crusoe, Core42, the U.S. Air Force, Department of Energy, Booking Holdings, Zoom, Boston Children's Hospital, and Pixar. The company reached approximately 1,150 employees globally by end of 2025 — a 26% year-over-year increase — with engineering concentrated in Tel Aviv and Haifa, and corporate functions in New York. VAST was named a Gartner Magic Quadrant Leader for File and Object Storage Platforms in October 2024, appeared on the 2025 CNBC Disruptor 50, and has earned the industry's highest verified Net Promoter Score of 84.
What does VAST Data offer?
VAST's product suite spans the full AI data stack — from all-flash distributed storage to vector databases, distributed Python compute, agentic AI orchestration, and cloud replication. The platform is marketed as the VAST AI Operating System.
- DataStore· Storage
- DataBase (vector + columnar)· Database
- DataEngine (distributed Python compute)· Compute
- DataSpace (metadata + policy namespace)· Data Management
- AgentEngine (multi-agent orchestration)· Agentic AI
- SyncEngine (universal cloud replication)· Cloud
- PolicyEngine (zero-trust AI governance)· AI Governance
- TuningEngine (continuous model learning)· AI/ML
- InsightEngine (NVIDIA co-developed)· AI/ML
- NVMe-oF All-Flash Storage· Infrastructure
- VAST AI Operating System· Platform
- Unified Namespace (NFS/SMB/S3/Kubernetes CSI)· Protocols
- VAST Federal (FedRAMP-adjacent)· Government
How does VAST Data make money?
VAST sells annual subscription software licenses priced on raw storage capacity and compute cores deployed, combined with hardware purchased by customers at manufacturer cost — a model that eliminates traditional vendor markup and drives gross margins approaching 90%.
The core commercial motion is a capacity-based subscription: customers pay approximately $30 per TB per year for the VAST software license on a 4–5 year contract term, billed upfront. Hardware is procured separately through certified OEM partners (HPE, Lenovo, Supermicro, Cisco, Avnet) at negotiated manufacturer cost — VAST negotiates bulk pricing on behalf of its entire customer base, then passes that manufacturer economics directly to buyers. This decouples software economics from hardware cycles and lets VAST retain near-100% of the software margin while eliminating the hardware margin that traditional storage vendors depend on. Average annual deal size exceeds $1 million; the top 100 new customers in 2024 averaged $1.2 million in annual commitments, and at least three customers have committed more than $100 million in total platform spend.
Contracts are long-dated (typically 5–7 years) and expand automatically with data growth, producing unusually low churn and strong net revenue retention. VAST reported net revenue retention above 300% in recent periods, meaning existing customers triple their spend on average over their contract lifetime. The CoreWeave $1.17 billion commercial agreement (November 2025) and the xAI Colossus deployment signal a growing hyperscaler segment that writes nine-figure, multi-year commitments — a structural shift from the enterprise storage procurement norms of the prior decade.
Growth is fueled by the platform expansion playbook: customers land on DataStore (storage), then expand into DataBase, DataEngine, AgentEngine, and the new PolicyEngine and TuningEngine services as AI workloads mature. This upsell path, combined with contractual capacity growth, drives a compound expansion effect. By end of 2025 VAST was generating more than $100 million in free cash flow per quarter, with positive operating margin and five consecutive years of positive free cash flow — an exceptional combination for a company growing at roughly 3x per year.
Who leads VAST Data?
VAST Data is led by its four co-founders — Renen Hallak (CEO), Alon Horev (CTO), Shachar Fienblit (Chief R&D Officer), and Jeff Denworth (product and commercial strategy) — augmented by an experienced executive team across finance, revenue, and marketing hired to prepare the company for a potential IPO.
- Renen HallakCo-Founder & CEO2016–presentFormer Head of R&D at XtremIO (acquired by EMC for ~$430M in 2012); architect of the DASE disaggregated storage concept and the primary public face of VAST at NVIDIA events and strategic customer briefings.
- Alon HorevCo-Founder & CTO2016–presentCore architect of VAST's NVMe-over-Fabrics storage fabric and the DASE architecture; leads all product engineering from Tel Aviv.
- Shachar FienblitCo-Founder & Chief R&D Officer2016–presentEx-Kaminario; drives platform expansion into DataBase, DataEngine, and the expanding VAST AI OS software domains including the 2026 PolicyEngine and TuningEngine releases.
- Jeff DenworthCo-Founder (Product & Commercial Strategy)2016–presentEx-CTERA Networks; leads strategic customer and partner relationships; the company's primary external product voice and evangelist at industry conferences.
- Michael WingPresident2022–presentOversees go-to-market, global partnerships, and field operations.
- Amy ShaperoChief Financial Officer2024–presentFormer CFO of Shopify (scaled revenue from $700M to $6B in five years), Betterment, and Sailthru; VAST's first-ever CFO, hired in August 2024 — the clearest public signal of IPO preparation.
- Rick ScurfieldChief Revenue Officer2024–presentVeteran enterprise technology sales executive leading global sales strategy and expansion into new industry verticals.
- Marianne BudnikChief Marketing Officer2023–presentFormer marketing leader at CrowdStrike, CyberArk, CA Technologies, and nearly a decade at EMC; leads brand, content, PR, and field marketing globally.
How do you contact VAST Data's leadership?
VAST Data's verified email format is first.last@vastdata.com (used by approximately 78% of employees per RocketReach). The leadership emails below follow this verified pattern; none are individually published and should be treated as format-derived outreach addresses, not confirmed personal inboxes.
first.last@vastdata.comHow much funding has VAST Data raised?
VAST Data has raised approximately $1.4 billion in equity across seven or more disclosed rounds since 2016. Its April 2026 Series F valued the company at $30 billion — more than three times its December 2023 Series E valuation of $9.1 billion and 25x its 2020 Series C valuation of $1.2 billion.
VAST launched with a $15 million seed/Series A in early 2016, led by 83North and Norwest Venture Partners, targeting NVMe all-flash storage for AI and HPC. In 2018, VAST raised a $25 million Series A1 with Dell Technologies Capital joining as a strategic investor. The company then raised a $40 million Series B in 2019 (Greenfield Partners leading, with 83North and Norwest). The $100 million Series C in April 2020 was led by Next47 at a $1.2 billion unicorn valuation, with co-investors including Goldman Sachs, Dell Technologies Capital, Mellanox Capital, Commonfund Capital, and Norwest.
The $83 million Series D in May 2021 was led by Tiger Global at a $3.7 billion valuation — tripling the Series C valuation in just 13 months — with NVIDIA joining as a strategic investor for the first time. After a period of heads-down execution, VAST closed a $118 million Series E in December 2023 led by Fidelity Management & Research Company at a $9.1 billion valuation, with co-investors including NEA, BOND Capital, Drive Capital, NVIDIA (returning), Dell, Goldman Sachs, Tiger Global, and others.
In April 2026, VAST closed its Series F: approximately $1 billion in combined primary and secondary capital (with over $500 million in secondaries allowing early investors and employees to realize liquidity). Drive Capital led the primary tranche, with Access Industries (family office of Len Blavatnik) as co-lead, and continued participation from Fidelity, NEA, and NVIDIA. The $30 billion valuation represents a 3.3x step-up from Series E in under 30 months, driven by revenue that grew from roughly $200 million ARR in early 2025 to approximately $2 billion ARR run-rate by end of 2025, combined with sustained positive free cash flow exceeding $100 million per quarter. The company has reported positive operating cash flow for over five consecutive years — extraordinary for an enterprise infrastructure company growing at this rate.
How did VAST Data get here?
VAST Data was founded in January 2016 by four XtremIO and Kaminario veterans and grew from a flash-storage upstart to a $30 billion AI Operating System company in under a decade.
- January 2016Company Founded with $15M SeedRenen Hallak, Alon Horev, Shachar Fienblit, and Jeff Denworth found VAST Data in New York, raising a $15M seed round from 83North and Norwest Venture Partners to build NVMe all-flash storage for AI and HPC.
- 2018Series A1 — Dell Technologies Capital JoinsRaises $25M Series A1 with Dell Technologies Capital joining as strategic investor, validating the enterprise storage opportunity.
- February 2019Series B — $40M, Greenfield LeadsCloses $40M Series B led by Greenfield Partners with continued participation from 83North and Norwest; product enters general availability.
- April 2020Series C at $1.2B Unicorn ValuationRaises $100M Series C led by Next47 at a $1.2B valuation, co-invested by Goldman Sachs, Dell Technologies Capital, Mellanox Capital, and others — VAST's first unicorn round.
- May 2021Series D — Tiger Global at $3.7B, NVIDIA JoinsRaises $83M led by Tiger Global at $3.7B valuation (3x step-up in 13 months); NVIDIA becomes a strategic investor for the first time, embedding VAST in the NVIDIA DGX ecosystem.
- December 2023Series E at $9.1B — Fidelity Leads, $1B in Cumulative BookingsCloses $118M Series E led by Fidelity Management & Research Company at $9.1B valuation (2.5x step-up from Series D); simultaneously announces $1B in cumulative software bookings and NVIDIA returns.
- November 2025$1.17B CoreWeave Commercial Agreement + Microsoft Azure PartnershipSigns landmark $1.17B multi-year deal with CoreWeave and announces VAST AI OS availability on Microsoft Azure, making VAST the primary data platform for two of the most important AI infrastructure players in the world.
- February 2026VAST Forward 2026 — PolicyEngine and TuningEngine UnveiledAt its annual conference, VAST unveils PolicyEngine (zero-trust AI agent governance) and TuningEngine (continuous model learning loops), completing the VAST AI OS closed-loop architecture.
- April 2026Series F at $30B — $4B+ Cumulative Bookings, ~$2B ARRRaises ~$1B Series F led by Drive Capital at $30B valuation; reports ~$2B ARR run-rate by end of 2025, $500M+ in committed ARR, $4B+ in cumulative bookings, and positive free cash flow exceeding $100M per quarter.
Who are VAST Data's competitors?
VAST competes in AI and HPC storage infrastructure against incumbent storage vendors, HPC-specialist challengers, and emerging architectural rivals.
- DDN (DataDirect Networks)Longest-standing HPC storage specialist; DDN's AI400X3 competes directly for GPU cluster storage with stronger entrenched presence in national labs and U.S. government HPC than VAST.
- WekaIO (WEKA)Cloud-native parallel file system optimized for NVIDIA DGX environments; competes on GPU-attached latency and cloud-burst capability but lacks VAST's unified database and compute layers.
- Pure StoragePublic storage incumbent (NYSE: PSTG); FlashBlade//S and FlashBlade//EXA are Pure's answers to VAST's disaggregated architecture, but Pure carries legacy product complexity and multi-tier overhead.
- NetAppLegacy enterprise NAS/SAN leader; competes via ONTAP and StorageGRID but is disadvantaged by multi-tier architecture and slower NVMe transition relative to VAST's all-flash native approach.
- IBM Storage Scale (GPFS)IBM's parallel file system dominant in government HPC and financial services; strong on reliability and compliance but ages poorly against cloud-native AI workloads and VAST's unified namespace.
- HammerspaceNewer challenger focused on global data orchestration across hybrid and multi-cloud environments; targets orchestration buyers rather than replacing underlying storage, making it more complementary than competitive in some accounts.
VAST Data — frequently asked questions
