NetApp plans to buy PEAK:AIO for its AI storage architecture
Founder Mark Klarzynski built three storage companies that were acquired before PEAK:AIO; NetApp wants its metadata technology alongside ONTAP.
By RuntimeWire Staff · Published
Primary source: Financial Times
Why it matters
NetApp says the planned integration will add PEAK:AIO's specialized metadata services and parallel namespace innovation to its ONTAP-based data architecture, while giving existing ONTAP customers an evolution path, according to [the acquisition announcement](https://markets.ft.com/data/announce/detail?dockey=600-202609250900BIZWIRE_USPRX____20260925_BW256928-1&ref=runtimewire). The purchase price remains undisclosed, and the announced scale targets have not been independently validated in production.
NetApp said on September 25th that it plans to acquire PEAK:AIO, the storage software company founded by serial storage entrepreneur Mark Klarzynski. The proposed deal would bring PEAK:AIO's metadata services and parallel file architecture into NetApp's ONTAP data platform, targeting a specific constraint in AI infrastructure: keeping data moving as GPU clusters and file counts grow.
Klarzynski has spent decades building storage technology for changing workloads. He formed Computer Design Group in 1991; Tandberg Data acquired it in 2005 for its BakStor technology. He later founded ID7 and SCST, both acquired by Fusion-io in 2013. PEAK:AIO's account of his career traces a repeat pattern: build specialized storage technology, then put it inside a larger vendor with the resources to distribute it. NetApp's planned acquisition would give that approach a much larger enterprise platform to work with.
The announcement quotes Klarzynski, PEAK:AIO's founder and CTO, saying the two companies share a belief that AI requires a new approach to data infrastructure. He described joining NetApp as a way to pair PEAK:AIO's innovation with the larger company's reach and customer trust. That is the proposed commercial bargain: NetApp gains specialist architecture, while PEAK:AIO's technology gets a route into an established storage business.
A targeted addition to ONTAP
NetApp says the architecture separates metadata services from data services so metadata can scale independently. Metadata tracks information about files and where they reside; at AI scale, coordinating access to vast numbers of files can become a bottleneck even when the storage hardware can deliver data quickly. PEAK:AIO's parallel NFS approach is intended to distribute that work across servers while preserving standard NFS access.
NetApp says the planned system is designed for trillions of files and multi-exabyte deployments. Those are target capabilities, not disclosed production results. The announcement does not provide customer benchmarks showing what the combined architecture can handle, or a timetable for integrating PEAK:AIO's technology with ONTAP. NetApp says the combination is intended to preserve ONTAP's resilience, security and operational characteristics while expanding metadata scale.
The purchase also fits a wider push by NetApp to add AI infrastructure capabilities through acquisitions. In July, NetApp announced its acquisition of DataPelago, another addition to that portfolio. PEAK:AIO would add a different piece: a storage architecture focused on parallel file access and metadata coordination. The strategic logic is clear from the product descriptions; how the components will be packaged and sold remains unspecified.
The storage market is crowded: vendors including WEKA, VAST Data and DDN already sell storage systems aimed at AI and high-performance computing. Separating metadata and data paths and offering parallel access are established approaches in this category. PEAK:AIO's case rests on its particular implementation, its NFS compatibility and NetApp's ability to deliver it to existing customers. The acquisition announcement offers no comparative benchmark against those competitors.
A young company built around an old problem
PEAK:AIO raised more than $6.8 million in seed funding in October 2025, according to the company's announcement. Pembroke VCT led the round with a reported $5 million investment; Praetura Ventures and an unnamed Silicon Valley investor also participated. The acquisition announcement does not disclose a purchase price, deal structure or expected closing date, so investors and employees cannot yet compare the proposed sale with the company's recent financing or assess its implied valuation.
The funding release described PEAK:AIO's next step as moving from single-node systems toward scaled-out AI infrastructure. Its later work with Los Alamos National Laboratory on Lattice, an open-source metadata server for parallel NFS, shows the technical direction behind the NetApp deal. The companies' June 2026 announcement described the system as dynamically scalable and said it could support up to 1,000 metadata servers. That is a company-and-partner description of the technology, not independent proof of performance at that scale.
NetApp's announcement positions the acquisition as a way to help customers scale shared storage alongside GPU clusters. Klarzynski's record gives the move a founder's logic: develop technology to solve a storage problem, then connect it to the operating scale and customer base needed for adoption. Whether PEAK:AIO's architecture becomes a distinct NetApp offering or disappears into ONTAP will determine how visible that technology remains. For now, the agreement is an intent to acquire, subject to customary closing conditions and regulatory approvals.