NetApp Buys Peak:AIO To Add Parallel NFS To AI Cloud Storage

NetApp is buying Peak:AIO to add parallel NFS and independently scaling metadata services to its ONTAP storage platform, Blocks & Files reported, turning a specialist UK storage startup into part of NetApp’s AI infrastructure roadmap.

The acquisition centers on a specific bottleneck in large AI clusters: storage has to keep pace with GPU growth without forcing customers to abandon the operational controls already built around existing arrays.

Peak:AIO brings software for a global file namespace and parallel NFS data access aimed at massively parallel workloads, while NetApp wants to fold those capabilities into the ONTAP base used by current enterprise customers.

Peak:AIO began in Manchester in 2021, with Mark Klarzynski serving as chief strategy officer and CTO and Eyal Lemberger leading AI and cyber security strategy.

Roger Cummings joined as US-based chief executive in May 2024.

Research ties shaped the technology before the deal, with Los Alamos National Laboratory and Carnegie Mellon University among the institutions connected to its development.

Its recent product history includes a 2RU DataServer parallel NFS NAS system using Dell server hardware, Solidigm NVMe storage and Nvidia CX7 Ethernet networking.

NetApp is framing the deal as an architecture change rather than a simple feature addition.

The planned design separates metadata from data so metadata services can scale independently, creating a storage foundation for AI infrastructure that may need to support trillions of files, exabyte-scale environments and highly concurrent workloads.

Peak:AIO has also discussed a CXL memory-based KV cache offload system and offers a flash-based PEAK:ARCHIVE product with up to 1.4 PB of capacity in a 2RU chassis.

NetApp chief executive George Kurian said AI clouds need shared storage that scales with growing GPU clusters while preserving resilience, security and operational simplicity.

The company argues that ONTAP and other long-standing storage designs need additional scale and concurrency for AI factories, AI clouds and newer data-heavy applications.

The practical target is lower data-related GPU stalls and a clearer path for shared storage to expand at the same pace as compute.

An ONTAP platform with Peak:AIO metadata capabilities and parallel namespace features would keep NetApp’s existing resilience and security model while adding a higher-throughput path for AI workloads.

For existing ONTAP customers, that evolution path is a central part of the pitch.

Rather than positioning Peak:AIO as a separate replacement stack, NetApp emphasized continuity for its installed base as AI workloads put more pressure on file services, namespace management and data movement.

The transaction still needs customary closing conditions and regulatory approvals.

Peak:AIO raised about $7 million in seed funding in October last year.

NetApp’s latest quarterly results included $410 million of free cash flow, so the purchase appears modest in financial size but strategically tied to AI cloud storage scaling.