NetApp moves to buy PEAK:AIO to scale AI storage metadata

25 Sep 2026 (ET): NetApp (NASDAQ: NTAP) announced its intent to acquire Manchester UK-based PEAK:AIO, a software-defined AI storage firm focused on next-generation metadata architecture and high-performance parallel file systems, to accelerate NetApp’s AI infrastructure roadmap.

FINANCE desk — a room of AI chips can sit idle when the shared storage cannot answer, fast enough, where each file lives. That lookup is the metadata, the index of the files. The bytes can be ready and the chips still wait. NetApp says it intends to buy PEAK:AIO, a Manchester company whose work is that index and a parallel way for many machines to share one set of files, and pair it with ONTAP, the storage software NetApp already sells. The page does not name a price or a closing date. The purchase is still an intent, waiting on the usual approvals. It is a tell on the AI build-out: the scarce piece being bought here is the index that lets the chips share one pile of files, not another chip and not another building.

What the wire says was announced, and the tense it uses. NetApp (NASDAQ: NTAP), which the lede calls the intelligent data infrastructure company, today announced its intent to acquire PEAK:AIO, which the lede calls a pioneer in next-generation metadata architecture and high-performance parallel file systems. The subhead calls it a planned acquisition. The last sentence before the legal block says the transaction remains subject to customary closing conditions and regulatory approvals. Intent, planned, and remains subject are the wire’s tense. This is not a sentence that the deal has closed. Pioneer is the lede’s label. This desk did not rank the market. NASDAQ: NTAP is the ticker on the lede. The page does not print a price, a revenue multiple, or a closing date. Do not invent them.

What the wire says the purchase is for. The planned acquisition is expected to accelerate NetApp’s AI infrastructure roadmap by augmenting metadata services and parallel namespace innovation designed to help AI clouds scale shared storage alongside growing GPU clusters. Expected, and designed to help, are the wire’s words. A GPU, here, is the chip that trains and runs an AI model. A cluster is a room of those chips working as one job. Shared storage is the file system those chips read from together. Those glosses are this desk’s. This desk did not time a cluster.

The problem paragraph, kept as what NetApp says it is building. As AI becomes embedded in every enterprise workload, the wire says, organizations are confronting a new challenge: traditional storage architectures were not designed for the scale, concurrency, and performance requirements of AI factories, AI clouds, and next-generation data-intensive applications. Unprecedented is the wire’s adjective on that scale. NetApp is building an architecture that disaggregates metadata from data and enables metadata services to scale independently, creating a foundation for AI infrastructure capable of supporting trillions of files, exabyte-scale environments, and massively parallel workloads. Metadata, here, is the index: the name of a file, where it lives, and who may open it. The bytes are the file itself. Disaggregates metadata from data means the lookup can grow apart from the disks that hold the bytes. An exabyte is about a billion gigabytes. Those glosses are this desk’s. Trillions of files, and exabyte-scale, are this paragraph’s. A later paragraph says multi-exabyte. Do not collapse exabyte-scale and multi-exabyte. Is building is this paragraph’s tense. A later paragraph says the architecture is intended to support. Do not collapse is building and is intended.

How the wire says the two products would sit together. The acquisition reinforces NetApp’s strategy to address the rapidly growing market for AI infrastructure. By combining PEAK:AIO’s specialized metadata innovations with NetApp ONTAP software, customers will benefit from a differentiated architecture that preserves the resilience, security, and operational maturity of ONTAP while introducing a new level of scalability for AI workloads. Will benefit is the wire’s tense. Resilience, security, and operational maturity are this paragraph’s three words. George Kurian’s quote, below, says resilience, security, and operational simplicity. A later paragraph says resiliency, security, and operational simplicity. Maturity and simplicity are different words. Resilience and resiliency are both on the page. Do not collapse the three sentences. ONTAP is NetApp’s data-management software. The about box also calls it an operating system. That gloss is this desk’s, from the about box. The wire prints a registered-mark on NetApp and on ONTAP. This desk did not open a product console.

George Kurian, as a quote, not as a cluster this desk timed. Kurian, Chief Executive Officer at NetApp, said: “AI clouds need high-performance shared storage that can scale alongside growing GPU clusters while maintaining the resilience, security, and operational simplicity organizations depend on.” He said: “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 maximize infrastructure efficiency and support AI at hyperscale cloud.” Hyperscale cloud is his phrase, as printed. Delivering, in that sentence, is his word for the vision. It is not a sentence that the deal has closed, and it is not a product this desk installed. The legal block, below, says NetApp makes no commitment to deliver the integrations the release describes.

Where the wire says the technology came from, and what the planned integration adds. PEAK:AIO’s technology originated from collaborations with leading research institutions, including Los Alamos National Laboratory and Carnegie Mellon University, and was designed to address some of the industry’s most demanding data-intensive computing challenges. Including is the wire’s word. The sentence does not say those labs sold the company, and it does not say they are the only collaborators. The planned integration will add specialized metadata services and parallel namespace innovation to NetApp’s ONTAP-based data architecture, while creating a clear evolution path for existing ONTAP customers. Will add, and planned, are the wire’s tense. A namespace, here, is the set of names a program uses to find files. Parallel namespace innovation is the wire’s phrase for that naming layer when many programs ask at once. Those glosses are this desk’s.

Mark Klarzynski, as a quote, not as a survey. Klarzynski, Chief Technology Officer and Founder at PEAK:AIO, said: “NetApp and PEAK:AIO connected through the shared belief that the AI era demands a fundamentally new approach to data infrastructure.” He said: “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.” He said: “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.” Joining is his word. Global leader is his label for NetApp. This desk did not rank NetApp, and it did not interview him.

What each side is said to bring, still the wire’s. PEAK:AIO’s metadata innovation extends metadata scale, delivers a global namespace, and enables parallel NFS-based access for massively parallel workloads. NetApp contributes trusted ONTAP expertise, global reach, and operational maturity to make AI-scale file infrastructure more practical and accessible. Together, they are expected to deliver new capabilities that simplify the deployment and operation of shared storage for increasingly large AI environments. A global namespace is one name for a file no matter which machine holds the bytes. NFS, network file system, is a standard way for many computers to share one set of files. Parallel NFS-based access means many of them can reach that set at once. The hed shortens that to pNFS. The wire does not print the letters pNFS. Those glosses are this desk’s. Expected to deliver is the wire’s tense.

The architecture sentence, kept apart from the earlier scale sentence. The architecture is intended to support trillions of files and multi-exabyte deployments while preserving the resiliency, security, and operational simplicity customers expect from ONTAP. Designed around the needs of large-scale AI environments, it combines scalable metadata services, standards-based parallel NFS access, and ONTAP as the resilient data foundation to help reduce data-related GPU stalls, improve infrastructure efficiency, and accelerate AI innovation. Intended, and to help reduce, are the wire’s limits. A stall, in data-related GPU stalls, is a chip waiting because the data is not ready. That gloss is this desk’s. Multi-exabyte is this paragraph’s. The earlier paragraph said exabyte-scale. Standards-based parallel NFS access is this paragraph’s phrase. The earlier paragraph said parallel NFS-based access. Do not collapse the two phrases. This desk did not measure a stall.

What the page does not sell as finished. The release says it contains forward-looking statements under the Private Securities Litigation Reform Act of 1995. A later block says the release discusses NetApp’s vision for future innovation, including the anticipated alignment and integration of PEAK:AIO’s technology with NetApp’s portfolio. It says that information is shared solely for informational purposes and should not be relied upon in making purchasing decisions. It says NetApp makes no commitment and has no obligation to develop or deliver any products, services, integrations, or related features described, including any capabilities resulting from the acquisition of PEAK:AIO. The development, release, and timing of any features remain at the sole discretion of NetApp. Strategy and possible future developments are subject to change without notice. Those sentences are the limit on the architecture above. They are not a second product. Tuck-in, in the hed, is this desk’s word for a larger company buying a smaller specialist. The wire does not print tuck-in. Undisclosed, in the hed, is this desk’s word for a page that does not print a price. The wire does not print the word undisclosed.

Who the about box says PEAK:AIO is, and where it says the software runs. PEAK:AIO is a Manchester UK-based software-defined AI storage company. Manchester UK-based is the box’s wording. It does not print a street. Its platform delivers high-performance AI storage from a single server to exabytes on any industry-standard hardware. From a single server to exabytes is the about box. It is not the multi-exabyte sentence in the architecture paragraph. The box says PEAK:AIO is deployed at Los Alamos National Laboratory, NHS AIDE, Oxford Robotics Institute, Carnegie Mellon University, University of Liverpool, University of Strathclyde MediForge Hub, and Zoological Society of London. Deployed at is the box’s word for that list. Los Alamos and Carnegie Mellon also appear in the collaborations sentence above. Do not collapse a collaboration origin and a deployment list into one claim the page does not print as one claim. The box points at https://peakaio.com. This desk did not open that site, and it did not visit a lab.

Who the about box says NetApp is, in short, and who to call. For more than three decades, the box says, NetApp has helped organizations navigate change, from enterprise storage to what it calls the intelligent era defined by data and AI. It says NetApp is the Intelligent Data Infrastructure company. The box says the data platform is built on ONTAP, called its leading data management software and OS. It also names automation through the AI Data Engine and AFX. The page does not spell out AFX. Do not invent an expansion. It says the platform is the only enterprise storage platform natively embedded in the world’s largest clouds. Only, and world’s largest, are the box’s claims. This desk did not audit the clouds. The site it names is www.netapp.com. The media contact is Kenya Hayes, NetApp, kenya.hayes@netapp.com. The investor contact is Kris Newton, NetApp, kris.newton@netapp.com. The mirror’s last line before the contacts points at the Business Wire URL this desk could not load.

Plain English for the rest of the card. NetApp, the storage company that trades as NTAP, said on 25 Sep 2026 that it intends to buy PEAK:AIO. PEAK:AIO is a software company in Manchester, in the UK, that builds storage for AI. The piece NetApp wants is the index of the files, called metadata, plus a way for many computers to share those files at once, called parallel NFS. ONTAP is the storage software NetApp already sells. The wire says the combined design is meant to handle trillions of files and deployments past an exabyte, and to cut the time AI chips spend waiting on data. The page does not say the price. It does not say the day the deal would close. It says the usual closing conditions and regulatory approvals still apply. It also says NetApp is not promising to ship the combined product.

PRIMARY here: NetApp’s 25 Sep 2026 Business Wire release 20260925256928, datelined San Jose — Tier A PRIMARY, the company’s own wire, read on the Financial Post mirror because the direct Business Wire page did not return. STATUS PRIMARY. The desk label on the chip is CONFIRMED because that is the catalog word for a primary page we can stand on. The intent sentence, the pioneer label, the planned-acquisition subhead, the closing-conditions sentence, the GPU-cluster purpose, the disaggregated-metadata paragraph, the trillions-of-files and exabyte-scale line, the ONTAP combination sentence, the Kurian quote, the Los Alamos and Carnegie Mellon origin sentence, the evolution-path sentence, the Klarzynski quote, the global-namespace and parallel-NFS sentence, the multi-exabyte and GPU-stalls sentence, the no-commitment block, the Manchester about box, the deployment list, and the two contacts are that wire’s. NOT claimed: a purchase price, a revenue multiple, a closing date, that the deal has closed, that delivering in the quote is a shipped product, that pioneer or global leader was surveyed, that exabyte-scale and multi-exabyte are the same phrase, that the collaborations list and the deployment list are the same claim, a definition of AFX the page does not print, a Business Wire header this desk did not load, a lab this desk visited, a stock tip, or investment advice. The card image is omitted. Distinct from the already-filed fed-schmid-ai-too-big-to-fail, nscale-pre-ipo-336b, and cognition-1b-arr. The catalog chip is FINANCE, not SIGNAL.