NetApp closes PEAK:AIO deal to scale file access for AI workloads
The acquisition adds metadata and parallel file-system expertise. NetApp targets much larger AI storage systems, without committing to delivery or timing.
Loading page…
The acquisition adds metadata and parallel file-system expertise. NetApp targets much larger AI storage systems, without committing to delivery or timing.
Listen to this story
NetApp’s purchase of PEAK:AIO gives it specialized metadata and parallel-file-system technology for a planned AI storage architecture, with an intended path into its ONTAP software. The design aims to scale shared file access as AI workloads grow, but support for trillions of files and zettabyte-scale deployments is a target, not a demonstrated capability. Integration details and timing remain undecided, and NetApp disclosed no financial terms.
NetApp completed the acquisition on November 2, 2026, targeting AI cloud operators and GPU rental providers.
The proposed architecture separates metadata from file data and plans parallel access through the NFS standard, without proprietary clients.
NetApp describes Novus as a potential capability, but says it is not committed to deliver the products or features in its product direction.
NetApp now owns PEAK:AIO, but the larger AI storage architecture it wants to build with the acquired technology remains a plan. The company announced the completed purchase on November 2, 2026, securing metadata and parallel file-system expertise intended to help keep large AI systems supplied with data.
The acquisition targets AI cloud operators and providers that rent access to graphics processing units, or GPUs. NetApp’s rationale is straightforward: those processors need a continuous supply of data, and storage bottlenecks can interrupt that flow. Financial terms were not disclosed in the announcement.
Metadata is the information used to keep track of files; a namespace provides a shared view of those files. NetApp’s planned additions include specialized metadata services, a global namespace and parallel access using the NFS file-sharing standard.
Chief Product Officer Syam Nair said the combined design separates metadata from the data itself, allowing performance to scale as more work happens at once. He also said it uses open standards rather than proprietary clients.
NetApp targets support for trillions of files and zettabyte-scale deployments while preserving resilience, security and operational simplicity. Those are anticipated benefits of the combined architecture, not demonstrated results.
The AI factory is only as productive as its data pipeline.
Syam Nair, NetApp chief product officer
NetApp plans to bring the acquired technology into ONTAP, its data-management software, creating an evolution path for existing customers. Nair said customers would be able to run more AI workloads on their existing infrastructure.
PEAK:AIO is a software-defined AI storage company based in Manchester, UK. NetApp says its platform runs on industry-standard hardware, from a single server to exabyte-scale systems. It names Los Alamos National Laboratory, Carnegie Mellon University, Oxford Robotics Institute and NHS AIDE among PEAK:AIO’s users.
Founder and Chief Technology Officer Mark Klarzynski said joining NetApp would pair that technology with the buyer’s reach, expertise and customer relationships.
NetApp connects the acquisition to Novus, describing it as a potential capability intended to simplify shared-storage deployment and operation. Its Statement of Product Direction explicitly says NetApp has no obligation to deliver the products, integrations or features described. Development, release and timing remain at the company’s discretion, and it warns customers not to base purchasing decisions on that vision.
Loading discussion...
Join the conversation
Explain when a supplier’s future plans deserve trust—or should be ignored.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.