VAST Data Reports $10 Billion in Customer Commitments From Leading AI Builders

Four customers have each committed more than $1 billion, as VAST arranges over 200 exabytes of flash supply through 2029.

[For more news, click here]

VAST Data has reported more than $10 billion in customer commitments for its software and supporting platform hardware during the first three quarters of its 2026 fiscal year. The commitments came from 14 of the more than 1,000 organizations the company serves, with each customer committing over $100 million. Four of them individually crossed the $1 billion mark.

The announcement, made on November 5, also highlights VAST's involvement in some of the largest AI infrastructure projects. According to the company, its systems now support nine of the top 10 frontier AI labs. Although the customers behind the commitments have not been named, the figures offer an indication of the spending taking place among organizations building AI systems at scale. The amounts represent customer commitments for software and supporting hardware, rather than revenue already recognized by VAST.

“Growth at this scale only happens when an entire industry is rebuilding at once,” said Renen Hallak, Founder and CEO of VAST Data. “That’s what the new AI stack is. It’s being built for models and agents, not people and applications, and data sits at the center of it. The companies building it at the largest scale, for people, agents and now robots, are choosing VAST to anchor it.” said VAST founder and CEO Renen Hallak

Alongside the customer commitments, VAST has structured agreements with its supply-chain partners to align more than 200 exabytes of flash capacity through 2029. The arrangements are intended to support the company's expansion plans and reflect the expected growth of its largest customers, while helping AI clouds, frontier labs, enterprises, and government organizations plan their infrastructure requirements.

Flash Supply Planned Through 2029

Expanding NAND flash production requires years of investment in fabrication facilities, making it difficult for manufacturers to respond quickly to rising requirements. VAST says the capacity its partners are delivering already represents a significant percentage of global NAND flash production for enterprise data centers. Its advance agreements are designed to give the company greater visibility into future availability and help it prepare for the storage requirements of upcoming deployments.

VAST attributes its understanding of future demand to its relationships with major AI developers. Its supply arrangements are also intended to benefit hundreds of clouds, rather than being directed exclusively toward its largest customers.

“Planning for hundreds of exabytes isn’t only about protecting our customers’ growth projections,” said Jeff Denworth, Co-Founder of VAST Data. “It’s about keeping flash available to hundreds of clouds, so that every AI builder, not just the largest, has the capacity it needs to build and compete.”

The infrastructure behind these deployments is VAST AI OS, which combines foundational data and compute services with agentic execution. The platform is designed to help organizations deploy AI agents, enable communication between them, reason over real-time data, and automate complex workflows at global scale. It is built on VAST's DASE architecture, a parallel distributed system designed to address performance, scale, simplicity, and resilience.

With its largest customers making substantial commitments and supply arrangements extending through 2029, VAST is preparing for several years of infrastructure expansion. The company's latest announcement brings its customer investment figures and long-term hardware planning into focus, showing the scale of resources it is arranging to support its growing AI infrastructure business.

Related Articles

VAST Data and NVIDIA Target the Infrastructure Challenges of Autonomous AI Agents

VAST Data and AMD Expand AI Infrastructure Alliance to Tackle the Industry's Inference Bottleneck

Exclusive: Matt Swalley on Why Storage Architecture Is the Next Major Bottleneck for AI Factories