Remote-First-Company | NEW YORK CITY and ABU DHABI – October 19, 2026 – VAST Data, the AI Operating System company, and Hyperfusion, the AI cloud and the modular AI data center company, today announced plans to integrate the VAST AI Operating System into Hyperfusion’s Sentinel modular AI infrastructure as its preferred data platform, creating a jointly developed offering designed to help customers bring sovereign, production AI inference online in weeks rather than waiting years for conventional data center capacity.
AI investment is increasingly colliding with a physical infrastructure constraint: GPUs can often be secured faster than the power, cooling and data center capacity needed to operate them. UAE-based Hyperfusion addresses that problem through factory-built modular infrastructure that can be deployed around available power. Its Sentinel platform packages a complete AI data center into a transportable 45-foot ISO High Cube unit, with a reference configuration of 96 GPUs, while HyperPod extends the same modular approach to larger, multi-megawatt deployments.
Every government and enterprise we speak to wants the same three things: capacity now, data that never leaves their control, and a platform that grows without starting over,” said Quentin Reyes, Founder & CEO. “Sentinel delivers the first two in a single container: a complete AI data center, commissioned in our factory and running on site in weeks, with confidential computing protecting models and data while in use. With VAST inside, every Sentinel extends the same data architecture as the last, so the tenth deployment scales as cleanly as the first.
Hyperfusion will take Sentinel to market as a repeatable modular architecture for government, enterprise, telecommunications, energy and other customers requiring rapid or sovereign AI deployment. With this partnership, Sentinel supports confidential computing as an additional security layer, protecting models and data while in use, and can operate fully air-gapped where required. Hyperfusion can deliver the infrastructure as a customer-owned system or operate it as a managed service. With the VAST AI OS as the preferred data platform, data stays in-country and under the customer’s control, while Sentinel and HyperPod deployments can share a single global namespace.
The model is particularly relevant to inference, where compute increasingly needs to sit closer to users, enterprise data or available power. Each Sentinel scaling unit is designed around 96 GPUs and approximately two to three petabytes of data capacity, creating a modular building block that can operate individually or as part of a larger environment, with each HyperPod scaling to approximately 20 petabytes. Hyperfusion also sees growing demand from multimodal models, computer vision and physical AI, where data requirements rise sharply alongside compute.
The AI infrastructure conversation has been dominated by how many GPUs you can secure, when increasingly the gating factor is where you can actually put them,” said Jeff Denworth, Co-Founder at VAST Data. “Power isn’t arriving in neat, hundred-megawatt blocks on the timetable AI demands. Hyperfusion gives customers a way to build around that reality by deploying compute wherever capacity exists. Our opportunity together is to make the data layer just as portable, so customers can stand up serious inference infrastructure quickly, without every new deployment becoming its own data island.
VAST and Hyperfusion will begin joint go-to-market activity, supporting customer introductions and opportunities across their respective markets while completing technical validation of VAST within Hyperfusion’s Sentinel and HyperPod architectures. The companies also plan to explore a common control plane for managing distributed Sentinel deployments across multiple sites.