VAST Data, Inc. designs software infrastructure to store, organize and use data at scale. Based in the United States, with headquarters in New York and research and development operations in Israel, the company serves large organizations, cloud service providers and high-performance computing specialists. Its positioning rests on a conviction: data architectures must evolve to power artificial intelligence applications without multiplying silos or data transfers.
From flash storage to a data platform
Founded in 2016 by Renen Hallak, Shachar Fienblit and Jeff Denworth, VAST Data emerged from stealth development in 2019. It first gained recognition with a storage architecture based entirely on flash memory, designed to accommodate data volumes traditionally stored on hard drives. The goal is to reconcile flash performance with the economic constraints of mass storage through hardware choices and software-based data reduction mechanisms.
This approach relies on an architecture known as “Disaggregated Shared Everything.” The computing resources that run storage services are separated from the media holding the data, while sharing access to them. This arrangement allows certain resources to scale independently and limits dependencies between capacity and processing power. In 2023, the announcement of the VAST Data Platform formalized the expansion of its offering beyond storage.
Bringing together storage, access and data use
The platform brings together file and object storage, database and processing capabilities. It aims to make information accessible for different uses through a common infrastructure, rather than imposing a succession of specialized systems. For technical teams, the priorities include reducing copies, simplifying administration and maintaining sufficient throughput when many servers access the same datasets simultaneously.
Use cases include artificial intelligence model training, analytics, scientific research and audiovisual production workflows. In these environments, storage performance partly determines how effectively graphics processing units are used. VAST Data works notably with the NVIDIA ecosystem to integrate its technology into AI infrastructure. Its business model centers on software, deployed with hardware equipment and integration partners, rather than solely on selling raw capacity.
What comes next?
The growth in AI adoption opens up a market for VAST Data, but also intensifies competition among storage specialists, data platforms and cloud providers. Its challenge will be to demonstrate that its proposed unification delivers lasting operational gains beyond advertised performance. Cost control, data governance and integration with existing systems will weigh on purchasing decisions. To continue growing, the company will also need to support the transition from AI experiments to services used daily, with their availability and security requirements.