TL;DR — Key Takeaways
– NetApp plans to acquire Peak:AIO to strengthen its storage capabilities for large-scale AI infrastructure.
– Peak:AIO separates metadata from underlying data, helping storage systems scale across billions or trillions of files.
– NetApp plans to integrate Peak:AIO technology with ONTAP to support multi-exabyte environments and parallel access to shared data.
NetApp plans to acquire Peak:AIO, adding storage technology created to keep massive GPU clusters supplied with data as AI infrastructure grows to trillions of files and multi-exabyte deployments.
GPU performance suffers if the storage systems feeding them can’t locate and deliver data quickly enough. As AI clusters add thousands of GPUs and handle huge numbers of files, storage architecture can become a limiting factor.
Peak:AIO addresses that problem by separating metadata from the underlying data, allowing metadata services to expand independently as workloads grow. Handling metadata efficiently becomes crucial when AI systems are working across billions or trillions of files.
NetApp plans to integrate Peak:AIO’s technology with ONTAP, its flagship data management and storage platform. The combined architecture is expected to provide a global namespace and parallel NFS-based access, enabling large numbers of computing resources to reach shared data simultaneously.
The goal is to support storage environments measured in multiple exabytes while reducing data-related GPU stalls. For enterprises spending heavily on AI accelerators, keeping processors working instead of waiting for data can improve efficiency.
The financial terms were not disclosed, and the transaction is subject to regulatory approval.
New Challenges for Data Storage
In the AI era, data storage faces new challenges because infrastructure is scaling along several dimensions at once. Enterprises are deploying more GPUs, working with larger datasets and managing far greater numbers of files.
Peak:AIO’s storage platform is built for these data-intensive computing environments, where low latency and high bandwidth are needed to drive GPU utilization. The company has received backing from investors including Pembroke VCT and Praetura Ventures.
Peak:AIO was founded in the U.K. in 2021 by storage industry entrepreneur Mark Klarzynski. Its technology has roots in work with Los Alamos National Laboratory and Carnegie Mellon University. Among the users of its technology are the U.K.’s National Health Service and Ministry of Defense.
For NetApp, the deal is the latest step in a rapid expansion of its AI and data infrastructure portfolio. In July, the company acquired DataPelago, which uses GPU acceleration for data processing at the storage layer. NetApp followed in August with its acquisition of JetStream Software, a provider of VMware disaster recovery and migration technology.
The three deals give NetApp capabilities across several parts of enterprise infrastructure. DataPelago adds accelerated data processing, JetStream brings migration and disaster recovery technology, and Peak:AIO is intended to strengthen storage performance for large-scale AI computing.
The strategy also boosts NetApp position in the infrastructure race, which has in a period of active growth. HPE has expanded its networking and hybrid IT portfolio through acquisitions including Juniper Networks, while IBM has purchased HashiCorp and Confluent to add capabilities around hybrid cloud and enterprise data.
