There is growing interest in hybrid data storage company VAST Data, founded a decade ago, which operates behind the scenes of the AI revolution. Over the weekend “Reuters” reported that the company is raising several billion dollars, including from Google and Nvidia, at a company valuation of up to $30 billion.
The company does not develop generative models like OpenAI’s ChatGPT or Google’s Gemini, but rather operates the infrastructure layer on which these models run. This is one of the less glamorous, but most critical areas in the AI era: storing, processing, managing, and retrieving vast amounts of data from diverse sources, at high speed and with great accuracy.
Moreover, unlike traditional systems that focus on structured data such as tables, VAST’s platform allows AI models to access unstructured data – text, images, video, sound, or free documents, and extract insights from them in real time.
In practice, this means that the model does not have to “wait” for the data to be converted, transferred between systems, or translated into a format it understands. The entire process is carried out on a single infrastructure, with full coordination between the storage components, databases, and servers that process the requests. This is a critical mechanism for intelligent agents and context-based search systems, but also for medical diagnostic tools, customer service, business intelligence, and more.
In a world where the speed of access to data and the ability to “understand” it immediately have become a differentiating factor between a mediocre model and a leading model, VAST offers a clear competitive advantage. The company provides not only a reduction in response times and infrastructure costs, but also a dramatic improvement in accuracy, resolution, and retrieval options of the data. The market is beginning to understand that such an infrastructure is not just technical support, but a fundamental condition for leaping to the next level in the field.
Prepare in advance for the revolution
At the core of VAST’s technology is a unique system called DASE – innovative architecture that unifies storage, data processing, databases and real-time analysis capabilities. Instead of several systems that transfer data between them slowly and in a complex way, as is the case with most rivals like NetApp, Pure Storage and Dell and Amazon’s cloud solutions, VAST centralizes the entire process in one place.
So while competitors offer separate infrastructure components, VAST becomes a comprehensive platform, almost like an operating system for the AI world. And proof: even before the astronomical investment on its part was announced, VAST and Nvidia, the chip giant leading the AI revolution, and which recently became the world’s most valuable company, have maintained a close technological relationship in recent years.
RELATED ARTICLES
VAST Data seeks to raise funding at $25b
Storage co Vast Data raising several billion dollars – report
was among the first in the world to adopt Nvidia’s BlueField technology – a dedicated chip (DPU) that transfers heavy processing tasks from the main processor to a separate component, which allows speeding up critical processes without burdening the system core.
This combination shortens response times, reduces server loads and saves significant energy and storage space. In practice, the result is a direct and fast connection between GPU servers and VAST’s data components. This is not only a matter of technological efficiency, but also economic potential. Customers running heavy models, such as cloud giants, research institutions or startups, can use VAST to provide accurate and fast AI services while dramatically reducing infrastructure costs.
With Nvidia, the cooperation does not stop at hardware. They recently launched a joint product called InsightEngine – a platform that demonstrates how storage and data infrastructure can become the engine of business AI applications. In doing so, VAST moves from offering general technology infrastructure to real edge solutions: from the toolbox of IT experts, to a product that customer service employees, research teams, analysts and managers can work with.
Both Google and the US government
Google sees VAST not only as an advanced technological supplier but also a genuine strategic partner. Last year, VAST’s system was integrated into Google’s cloud service, so that the company’s customers can use it to run AI models, analyze large amounts of data, and perform a host of advanced operations.
One of the most outstanding capabilities offered by VAST’s system is “vector search,” a smart method for locating data based not on exact words but on understanding meaning. Unlike regular search, where you have to type in exactly what you’re looking for, vector search also identifies similar ideas, different wordings, or close concepts, and “returns” results that match the context, not just the wording.
This is precisely the technology that allows chatbots to remember what was said in previous conversations, offer more accurate answers, and behave in an almost human manner.
So while other systems have to go through several layers of software to perform such a search, VAST’s platform can perform it “on the spot”, quickly, from a single system, without delays and dependence on external factors.
It can also locate the data wherever it is stored, whether on the organization’s computer, in the cloud or on other servers. This saves Google complexity. It can offer customers one simple, fast and smart solution. For VAST, this is confirmation that its approach is already proving itself even with top-tier global players.
It is not just the tech giants. The US government also recognizes the potential. Last July, VAST was chosen to take part in one of the most important scientific projects in the US today: a new supercomputer called Doudna, which is being developed by the US Department of Energy.
The project is led by the NERSC laboratory, one of the world’s most important scientific computing centers. As part of the project, IBM will be responsible for the traditional computing parts, while VAST has been awarded the sensitive and advanced role of managing the entire AI layer. This means that its system will handle not only the storage of huge amounts of data, but also management, transfer, and protection of data in the system, down to the level of the operating system itself. This is a reserved and restricted area, in which only a limited number of companies are authorized to participate.
Impressive business performance
Interest in the company is not only based on technological depth, but also impressive business performance. Earlier this year, it reported annual recurring revenue (ARR) of $200 million, and it expects to reach $600 million by the end of 2026, a figure that places it among the fastest-growing companies in the field of AI infrastructure. Although it has raised less than $400 million since its inception (excluding the current round of funding) – a relatively modest amount in the field – it has managed to maintain a high growth rate, early profitability, and a select list of customers that includes Elon Musk’s xAI, NASA, CoreWeave, and defense, medical, and financial corporations. The business model is simple and efficient. Selling software infrastructure licensing, with almost no marketing expenses, and with a team of only about 1,000 employees, including 500 in Tel Aviv.
VAST was founded in 2015 by Renen Hallak, Shachar Fienblit, and Jeff Danworth. They have led the company through a decade in which many data storage startups have disappeared from the market. While others competed for headlines, VAST focused on performance, and now, as demand for AI infrastructures increases, it finds itself well positioned. For investors, in an era where data has become the central engine of the economy, whoever holds the key to it also holds the future.
Published by Globes, Israel business news – en.globes.co.il – on August 4, 2025.
ยฉ Copyright of Globes Publisher Itonut (1983) Ltd., 2025.