AI CLOUD CEO INSIGHT 2026 Sep 30, 2026
The positioning. In the CRN interview, Kurian said NetApp’s software “brings together unified storage, security, data protection, metadata, and movement across data centers and clouds, allowing customers to build AI-ready data in place while retaining control.” His framing is that the differentiation is no longer the array but the software layer above it: “We have long ago progressed beyond just having the most capable storage on the planet. We have a suite of software that really allows people to use their data effectively.”
The money quote. Asked about the field of storage vendors now embracing cloud, Kurian drew a line: “Everyone else is doing what we did 10 years ago. We are the only natively-integrated cloud platform. We are bringing other cloud providers like Oracle Cloud and SAP Cloud into our ecosystem, so that now you can get native offerings in the Oracle Cloud, just like you can do in the other hyperscalers.” That is the shortest accurate version of NetApp’s multicloud argument: ONTAP-as-a-service in the hyperscalers plus first-party integrations, rather than a gateway bolted in front of someone else’s object store. The platform mechanics of that model are covered in our Cloud Volumes ONTAP reference.
The AI-factory bets. Kurian tied the planned Peak:AIO acquisition and the DataPelago technology to one outcome: shortening the path between storage and GPU compute. His claim — “Peak:AIO gives us the ability to deliver to customers the highest performance, most scalable infrastructure for AI factories, model builders, and hyperscalers, and NetApp with Peak:AIO is an order of magnitude better than anybody else” — describes the market NetApp is aiming at. DataPelago, he said, “is an acceleration technology that sits at the layer above storage to make processing frameworks much more optimized.” He named Novus as the AI-factory offering that pairs with it, and said the combination “can allow you to feed at line rate.”
Why the storage layer matters here. The operational consequence for ONTAP admins is that the AI-era pitch is not about raw IOPS but about data movement discipline. Kurian’s argument on cost and placement: “Cloud is not inherently cheaper or inherently more expensive. It is really tied to how you use it. It’s the same with on-premises.” He pointed to two existing ONTAP behaviours as the differentiators — automated tiering of aging data from flash to capacity disk “without our end users needing to do anything,” which he called “unique in the market,” and cloud bursting, citing “extraordinarily fast uptake of the work that we did with Amazon around S3 access points.” For teams standardising a hybrid model, the trade-offs are laid out in our on-premises versus cloud guide and the hybrid cloud hub.
Hybrid, not cloud-only. Kurian rejected the idea that multicloud dilutes on-premises value: “It actually enhances the value of our on-premises storage.” He cited customers reserving cloud instances over concerns about hardware availability and bursting on-prem data into cloud applications, and repeated a figure NetApp has shared with investors — “approximately 50 percent of our cloud customers have never bought NetApp on-premises.” On SAP, he confirmed a strategic collaboration that “starting at the customer premises where a lot of SAP deployments are today” is intended to become a hybrid environment over time.
What the interview does not say. No product was announced, no version or CVE is referenced, and the Peak:AIO deal is described as planned — so treat this as strategy and roadmap signal, not a change to plan into a maintenance window. The read-through that is actionable: NetApp is putting its cloud and AI narrative on native integrations (Oracle Cloud, SAP Cloud) and on the storage-to-GPU data path, which means the ONTAP skills that matter next are data-movement and tiering policy, not just provisioning.
Read the original (CRN) · Hybrid cloud hub · Cloud Volumes ONTAP reference · On-prem vs cloud