- Vast Data closed a $1 billion funding round on April 22, 2026, including a secondary offering that enables existing shareholders to sell stakes.
- The round values the company at $30 billion, approximately triple its valuation from its most recent prior fundraise.
- Nvidia remains a backer, reinforcing Vast Data’s position within the GPU maker’s AI infrastructure ecosystem.
- The company’s VAST Data Platform uses a disaggregated, NVMe-based architecture designed for large-scale AI training and inference workloads.
What Happened
Vast Data, the New York-based AI storage infrastructure company, raised $1 billion in a new funding round announced April 22, 2026, according to Bloomberg. The round values the company at $30 billion — roughly triple its valuation from its prior capital raise — and includes a secondary component enabling early investors and employees to sell existing shares. Nvidia, a strategic investor in prior rounds, remained among the company’s backers.
Why It Matters
The valuation increase reflects sustained enterprise demand for purpose-built AI storage infrastructure as organizations scale GPU clusters for model training and inference. Vast Data competes with storage incumbents including Pure Storage and NetApp, as well as AI-focused entrants like WEKA, all of which have repositioned their products around the high-bandwidth, low-latency data access requirements of GPU workloads. The secondary offering structure indicates investors anticipate a meaningful runway before a public exit rather than an imminent IPO.
Technical Details
Vast Data’s core product, the VAST Data Platform, uses a disaggregated storage architecture that decouples compute from storage, allowing both to scale independently — a departure from traditional scale-out NAS designs that tie the two together. The platform relies on QLC NAND flash accessed over NVMe over Fabrics (NVMe-oF), enabling multi-gigabyte-per-second sequential throughput and sub-millisecond latency for GPU clusters during training runs. According to the company’s published technical documentation, the architecture sustains millions of IOPS across petabyte-scale datasets without performance degradation as cluster size grows. CEO Renen Hallak has described the platform as a unified data layer designed to serve training, inference, and analytics workloads under a single namespace, eliminating the need for separate storage silos per workload type.
Who’s Affected
Enterprises running AI workloads on Nvidia DGX or HGX GPU clusters are the primary customer segment, as Vast Data’s platform is engineered as a high-performance storage backend for GPU-dense environments. Financial institutions, research organizations, and large technology companies already deploying the VAST Data Platform stand to benefit from the additional R&D investment the round enables. Competitors in the AI storage market — including Pure Storage’s FlashBlade, WEKA’s parallel file system, and IBM’s Spectrum Scale — face a better-capitalized rival as Vast Data scales its engineering and go-to-market operations.
What’s Next
Vast Data has not publicly specified how primary proceeds will be allocated, though prior fundraises directed capital toward expanding software capabilities and geographic sales coverage. The secondary offering component indicates the company is providing early-investor liquidity ahead of an eventual public market exit. Bloomberg did not report a timeline for an IPO or acquisition.
