NetApp closed its fiscal 2026 with numbers that would have been hard to imagine for a storage incumbent two years ago. Revenue for the April quarter came in at $1.948 billion, up 12 percent year on year and just past the top of guidance. Full year revenue hit $4.93 billion, the highest in the company's history. All flash arrays, the C series and ASA lines that NetApp has been pushing as the foundation for AI workloads, generated $1.2 billion in the quarter, up 18 percent. Public cloud revenue, which includes Azure NetApp Files, Amazon FSx for NetApp ONTAP and Google Cloud NetApp Volumes, reached a record $182 million, up 11 percent and around 9.3 percent of the top line. Product revenue alone hit $966 million, up 14 percent, with the multi year Google Distributed Cloud agreement cited as a meaningful contributor.
The number that should grab the attention of any data platform team is the one CEO George Kurian highlighted on the call: roughly 500 AI and data preparation wins in Q4 alone, compared to around 400 for the entire prior fiscal year. Even allowing for some elasticity in what counts as an AI win, the trajectory is unmistakable. Enterprises are not just buying GPUs, they are also buying the high performance, governed storage that feeds them. NetApp's pitch is that you can use the same ONTAP operating system for your transactional NAS, your VMware estate, your Kubernetes persistent volumes and your AI training corpus, with snapshots, replication and RBAC consistent across all of them. Kurian also noted that in less demanding AI environments customers are coming back to hybrid flash, a tier NetApp has kept investing in.
What the Google Distributed Cloud deal actually buys
The Google Cloud deal is the other strategic signal. NetApp announced a multi year agreement to deliver AI ready infrastructure to Google Distributed Cloud, the on prem and edge version of Google Cloud that customers use when they need to run Gemini or other Google services inside their own datacenters. ONTAP becoming a first class storage option for Google Distributed Cloud means that European banks, public sector buyers and regulated retailers who want to keep model weights and customer data on premises can do so without rebuilding their storage stack. For us as operators, that is a real procurement option, not a slideware partnership.
It is worth being concrete about what the Google Distributed Cloud integration actually buys. GDC ships as a Google managed rack or set of racks running the same control plane as Google Cloud, including Vertex AI and Gemini endpoints, but inside the customer's facility. Until now, buyers were stuck either using GDC's own local storage or bolting a NAS outside the GDC trust boundary. With ONTAP inside the reference architecture, the FlexClone and SnapMirror primitives that data teams already use for database refreshes and Kubernetes volume cloning extend to model checkpoints, retrieval indexes and fine tuning datasets. For a regulated buyer running sovereign Gemini inference, the practical benefit is that DR, backup and audit tooling do not have to be reinvented for the AI estate.
Margins held up despite NAND price increases. Gross margin expanded to 70.1 percent from 68.9 percent a year ago, helped by the public cloud mix shift. Operating cash flow grew to $950 million from $675 million, free cash flow to $900 million from $640 million, and EPS to $2.43 from $1.93. Cash and equivalents stood at $2.1 billion against $2.5 billion of gross debt, leaving roughly $1.1 billion of net cash. Kurian noted that NetApp passed through some NAND cost increases to customers and that the increases were accepted, which tells us the pricing power is still on the storage vendor's side in 2026. That is not great news for data platform budgets, where flash density assumptions made in 2024 are now being revisited as DRAM and NAND tightness extends.
Where NetApp sits in the storage three-way
The competitive picture matters. NetApp's 12 percent growth came in ahead of Dell's storage segment, which grew around 8 percent in its most recent comparable quarter, but trailed Pure Storage, which has been printing growth above 30 percent on the back of its E family and hyperscaler design wins. VAST Data, still private, continues to take share at the very high end where exabyte AI training clusters are the norm. The read for buyers is that the primary storage market has three live narratives at once: NetApp selling consolidation across virtualization, containers and AI on one ONTAP fleet; Pure and VAST selling dense QLC flash tuned for GPU feeding; and Dell selling breadth across PowerStore, PowerScale and PowerMax with deep server bundling. We expect 2027 bake offs to come down to total cost per usable petabyte and how cleanly each platform handles mixed protocol AI pipelines.
Guidance is the most aggressive part of the print. NetApp called for Q1 FY2027 revenue of $1.825 billion at the midpoint, roughly 17 percent growth year on year, and full year FY2027 revenue of $7.45 billion. That implies acceleration from the 5 percent growth NetApp posted for FY2026 as a whole, and it bakes in the assumption that AI infrastructure spending continues at the current pace. William Blair analyst Jason Ader called out strong traction in high performance flash for AI as well as broader datacenter modernisation and consolidation, suggesting the demand is not purely AI driven.
What this means for 2027 storage budgets
For data engineering leaders the operator takeaways are concrete. First, if your storage refresh has been on hold waiting to see how the AI architecture shakes out, the question is no longer whether to refresh but whether to consolidate. NetApp's argument that one ONTAP fleet can serve VMware, Kubernetes, file shares and AI training is increasingly credible, and Pure Storage, Dell PowerStore and VAST are all making similar pitches. Second, plan for NAND cost pass through in your 2027 budget. The supply tightness is real, and the vendors are not absorbing it. Third, if you are evaluating Google Distributed Cloud for sovereign Gemini deployments, ONTAP is now in the reference architecture. That is a less risky path than rolling your own storage layer under a Google managed control plane.
The bigger picture is that the storage market is being reshaped by AI in a way that looks different from the lakehouse story. Snowflake and Databricks are competing for analytic and AI workloads at the platform layer, with object storage underneath. NetApp, VAST, Pure and Dell are competing for the primary data layer that sits directly under GPUs and under enterprise applications. Both layers are growing, and the boundary between them is moving. Iceberg tables on S3 are an analytic pattern; checkpoints, KV caches and embeddings on parallel file systems are an AI pattern. Most enterprises will end up running both, and the architectural choice is whether to let them stay separate or to push for a single namespace. NetApp's bet is on the latter, and the numbers say that bet is paying off.
