NetApp buys UK's Peak:AIO to bolt parallel NFS onto ONTAP for AI
What's the deal? NetApp is acquiring Peak:AIO, a Manchester, UK-based software-defined all-flash storage startup, for an undisclosed sum. The target's software provides a global file namespace and parallel NFS data access for massively parallel AI workloads. NetApp announced the deal in September 2026.
Who is Peak:AIO? Founded in 2021 by Mark Klarzynski and Eyal Lemberger, the startup builds AI-focused storage software drawing on collaborations with Los Alamos National Laboratory and Carnegie Mellon University. US-based Roger Cummings joined as chief executive officer in May 2024. It raised around $7 million in seed funding in October 2025.
Why now? NetApp says traditional architectures like its ONTAP platform "were not designed for the unprecedented scale, concurrency, and performance requirements of AI factories, AI clouds, and next-generation data-intensive applications." Peak:AIO's technology plugs that gap.
What's the endgame? NetApp wants an architecture that separates metadata from data, letting metadata services scale independently. The goal: support trillions of files and exabyte-scale environments while cutting the data stalls that idle GPUs.
"AI clouds need high-performance shared storage that can scale alongside growing GPU clusters," said George Kurian, NetApp's chief executive officer. The plan pairs Peak:AIO's metadata services with "the proven foundation of ONTAP," he added.
What changes for customers? NetApp says existing ONTAP users will get a clear evolution path, with the acquired parallel namespace features layered onto the platform's existing resilience and security. The deal remains subject to customary closing conditions and regulatory approvals.
"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," Klarzynski said.
The signal: The purchase is small — likely immaterial against NetApp's $410 million in free cash flow last quarter — but strategically pointed. As storage vendors race to feed ever-larger GPU clusters, buying a specialist for parallel access and metadata scaling is a faster route to an AI-ready platform than building one.
Read more: blocksandfiles.com
Image credit: Generated with Gemini
