NetApp moves to buy PEAK:AIO to scale ONTAP storage for trillion-file AI clouds
NetApp has announced its intent to acquire PEAK:AIO, a developer of metadata architecture and high-performance parallel file systems, in a deal designed to help AI clouds expand shared storage at the same pace as their GPU clusters. The company did not disclose financial terms, and the transaction remains subject to customary closing conditions and regulatory approvals.
The deal targets a problem that has grown with enterprise AI adoption. Storage architectures built for conventional workloads were never designed for the scale and concurrency now demanded by AI factories, AI clouds and data-intensive applications. NetApp's answer is an architecture that separates metadata from data, allowing metadata services to scale independently, with the company stating that the design is intended to support trillions of files and multi-exabyte deployments.
"AI clouds need high-performance shared storage that can scale alongside growing GPU clusters while maintaining the resilience, security, and operational simplicity organisations depend on," said George Kurian, Chief Executive Officer at NetApp.
Research-lab origins give NetApp a head start on metadata at extreme scale
PEAK:AIO's technology grew out of work with Los Alamos National Laboratory and Carnegie Mellon University, where it was developed to handle some of the most demanding data-intensive computing problems in research. Its contribution to the combined platform is an extended metadata layer, a global namespace and parallel NFS-based access for massively parallel workloads, which NetApp plans to integrate into its ONTAP data architecture while offering existing ONTAP customers a clear path to adopt it.
"NetApp and PEAK:AIO connected through the shared belief that the AI era demands a fundamentally new approach to data infrastructure," said Mark Klarzynski, Chief Technology Officer and Founder at PEAK:AIO. "What excites us most about joining NetApp is the opportunity to pair our culture of innovation with the reach, scale, and customer trust of a global leader."
Kurian positioned the acquisition as the next step in the company's AI infrastructure roadmap. "With PEAK:AIO, we are delivering on our vision for a new generation of AI infrastructure that combines scalable metadata services with the proven foundation of ONTAP to help customers maximise infrastructure efficiency and support AI at hyperscale cloud," he said.
Idle GPUs are turning storage into a strategic infrastructure decision
The significance of the deal lies in where AI spending is being wasted. Organisations have poured capital into GPU capacity, yet those accelerators sit idle whenever storage cannot deliver data fast enough, and NetApp is pitching the combined architecture as a way to reduce those data-related GPU stalls and improve overall infrastructure efficiency. As clusters grow from hundreds to tens of thousands of GPUs, metadata operations such as opening, listing and locating files become a bottleneck long before raw capacity does.
The move also sharpens competition in a market where specialist vendors have built their businesses around parallel file systems and disaggregated designs for AI training and inference. NetApp's bet is that enterprises and AI cloud operators will favour standards-based parallel NFS access running on a platform their teams already know, rather than adopting new proprietary storage stacks to reach AI scale. For the large installed base of ONTAP customers worldwide, that could mean expanding into AI workloads without abandoning the data management, security and operational tooling they already rely on.
"Together with NetApp, we have an opportunity to help shape that future and deliver technologies that make managing data at a massive scale both practical and accessible," Klarzynski added.
