Vast Data positions the data layer for AI scale – News

The AI arms race has centered on compute: Who has the most graphics processing units, the fastest chips and the biggest clusters? But a different pressure point is emerging as enterprises move AI from pilot programs into continuous, production-grade operations. Vast Data Inc. has built its strategy around that gap, offering a data infrastructure platform designed to serve as the operating layer between enterprise datasets and the AI workloads that depend on them.

That ambition is backed by measurable momentum. Vast crossed $2 billion in cumulative software bookings in under six years, has maintained positive cash flow for 12 consecutive quarters and holds a gross margin approaching 90%, according to Paul Nashawaty, principal analyst at theCUBE Research.

“Vast Data is really carving out a unique spot in the cloud-native world,” he said. “They’re not just another storage vendor; they’re going after the role of a cloud-native AI infrastructure platform.”

Partnerships with Nvidia Corp., Microsoft Corp., CoreWeave Inc., Google Cloud and Solidigm, a trademark of SK Hynix NAND Product Solutions Corp., signal that major infrastructure players increasingly view the platform as foundational to AI at scale, according to Dave Vellante, chief analyst at theCUBE Research.

“We are witnessing the evolution of the ‘AI Factory,’ where the focus is shifting from experimental model training to continuous, high-scale inference,” he said. “For the enterprise, the challenge is increasingly around data velocity and architectural efficiency. Vast Data is at the center of this transition, providing the high-performance foundation required to run agent-based systems in production.”

That shift will be on display at Vast Forward. Join theCUBE, News Media’s livestreaming studio, on Feb. 25 for exclusive coverage of Vast’s evolving platform architecture, the operational demands of agentic AI and how the company’s expanding partner ecosystem supports production-scale deployments. (* Disclosure below.)

Inside Vast Data’s AI operating system

Vast’s platform comprises four integrated layers — DataStore, DataSpace, DataBase and DataEngine — each addressing a different stage of the AI data pipeline, from storage and cataloging to real-time analytics and distributed processing. The Disaggregated Shared-Everything Architecture, or DASE, separates compute logic from storage, enabling organizations to scale capacity and processing power independently. The approach positions the platform as an alternative to traditional data lake and lakehouse models, with an architecture designed to manage AI workloads end-to-end.

“In our opinion, success in AI requires more than just access to GPUs,” Vellante said. “Keeping those valuable resources will be just as important. Vast is playing directly into that reality with an architecture and an operating model built around utilization.”

Vast Forward is expected to build on themes introduced at the company’s Cosmos event and reinforced at KubeCon + CloudNativeCon NA, where Vast detailed how its platform embeds compute directly into the data layer. That capability enables AI models — including embedding models for video and audio — to operate across cloud and on-premises environments with Kubernetes as the orchestration layer. Product updates around DataEngine and DataSpace are among the areas to watch, along with expanded support for edge processing and autonomous agent workflows, according to Vellante.

“Especially in a market where flash and memory efficiency are becoming tactical mandates, Vast’s ability to optimize the data layer is a critical differentiator for any organization moving toward an always-on AI reality,” he said. “Our research indicates the industry’s memory constraints are forcing efficiency to the forefront — and Vast is leaning in tactically with flash efficiency programs designed to do more work per byte and per watt.”

An ecosystem for production-scale AI

Vast’s partner strategy spans cloud, compute and storage, extending earlier collaborations with Nvidia, Microsoft and CoreWeave while adding a recent alliance with Google Cloud. The Microsoft partnership, announced in late 2025, brings the Vast AI OS to Azure, giving enterprise customers access to tools such as InsightEngine for real-time data analysis and AgentEngine for orchestrating autonomous AI workflows across hybrid cloud environments, according to Jeff Denworth, co-founder of Vast Data.

“The objective is to basically bring all of the goodness of what we call the Vast AI Operating System to Microsoft customers,” he said. “We’ve been collaborating with them for, I’d say, the better part of two years now. The objective is to basically open up data access to all of the customer’s data sets for all the different compute platforms that Microsoft offers.”

On the compute infrastructure side, Vast and CoreWeave formalized a $1.17 billion commercial agreement that positions the Vast AI OS as a primary data foundation across CoreWeave’s network of GPU-dense cloud data centers. The deal allows CoreWeave to deploy Vast’s platform on demand, giving joint customers low-latency access to massive datasets for AI training and inference without the bottlenecks associated with conventional storage architectures.

“Our deep integration with CoreWeave is the result of a long-term commitment to working side by side at both the business and technical level,” said Renen Hallak, founder and chief executive officer of Vast. “By aligning our roadmaps, we are delivering an AI platform that organizations cannot find anywhere else in the market.”

Solidigm, another key Vast partner expected to feature prominently at Vast Forward, adds a storage economics dimension. A joint whitepaper from the two companies found that all-flash architectures deliver a 58.9% lower total cost of ownership than traditional hard-disk-drive tiers, challenging long-held assumptions about the need for tiered storage, according to the paper.

“We believe the timing of Vast’s customer event is fortuitous because it spotlights a real inflection,” Vellante said. “Specifically, AI infrastructure is moving from experimentation to production, and efficiency — especially in memory and flash — has become a first-order key performance indicator.”

TheCUBE event livestream

Don’t miss theCUBE’s coverage of Vast Forward on Feb. 25. Plus, you can watch theCUBE’s event coverage on demand after the event.

How to watch theCUBE interviews

We offer you various ways to watch theCUBE’s coverage of Vast Forward, including theCUBE’s dedicated website and YouTube channel. You can also get all the coverage from this year’s events on  News.

TheCUBE podcasts

News’s “theCUBE Pod” is available on Apple Podcasts, Spotify and YouTube, which you can enjoy while on the go. During each podcast, News’s John Furrier and Dave Vellante unpack the biggest trends in enterprise tech — from AI and cloud to regulation and workplace culture — with exclusive context and analysis.

News also produces our weekly “Breaking Analysis” program, where Dave Vellante examines the top stories in enterprise tech, combining insights from theCUBE with spending data from Enterprise Technology Research, available on Apple Podcasts, Spotify and YouTube.

Guests

During Vast Forward, theCUBE will feature interviews with executive leaders at Vast, Nvidia, Solidigm, CoreWeave and more. Discussions will focus on the impact of agentic AI on operations and how platform architectures are evolving to address these challenges.

(* Disclosure: TheCUBE is a media partner for Vast Forward. Sponsors of theCUBE’s coverage, including presenting sponsor Solidigm, do not have editorial control over content on theCUBE or News.)

Image: News