NetApp claims its new Novus file system delivers more than 100 TB/sec of scale-out bandwidth, a figure it positions as the fastest storage on the planet and the centerpiece of a broad product push announced at its annual Insight event aimed squarely at the AI factory market.

Novus disaggregates metadata and data servers, separating metadata from the data path so performance, capacity, and concurrency scale independently under a single namespace using industry-standard NFS. The timing is notable: NetApp acquired parallel NFS software supplier Peak:AIO just days before the announcement, a tuck-in deal that appears to fill a direct technology gap.

The initial Novus release combines Novus Data Director metadata software running on qualified Supermicro server infrastructure with ONTAP data services delivered through AFF A90 systems. NetApp says Novus is built to exceed 100 TB/sec bandwidth and support high utilization of hundreds of thousands of GPUs while preserving ONTAP enterprise data services. The launch messaging at one point claims millions of GPUs can be kept continuously fed, with zettabyte-scale support — a claim NetApp itself downgrades to hundreds of thousands of GPUs elsewhere in the announcement material.

The performance bar Novus must clear is well documented in NetApp's own published results. ONTAP on AFF systems (A800, A90 and similar) has supported GPU Direct Storage via NFS over RDMA since ONTAP 9.12.1, with published results of roughly 171 GiB/s on A800 and 351 GiB/s on an A90 cluster. NetApp AFX, the disaggregated ONTAP variant for AI certified for Nvidia DGX SuperPOD, has posted 457 GiB/s sustained across eight nodes. Novus is expected to comfortably exceed those numbers.

Asked whether Novus supports GPU Direct Storage via NFS over RDMA, pNFS and session trunking, Arindam Banerjee, NetApp's Chief Platform and Technology Officer, said: "NetApp Novus is built on industry standard pNFS protocols so it is not limited to any particular GPU architecture."

NetApp positions Novus as a direct competitor to high-performance AI factory storage specialists DDN, Everspin competitor Everpure, WEKA and VAST Data. Syam Nair, NetApp's Chief Product Officer, was blunt about the ambition: "HPC-era storage alone cannot serve AI Factory scale — the industry needs a true architecture built for it from the ground up. AI factories have long been searching for infrastructure that can scale at AI speed. Leveraging the power of the NetApp Platform, Novus delivers 100 TB/sec of unlimited scale-out performance, leapfrogging the market to serve as the new high-performance, AI-scale storage architecture that unlocks the performance of millions of GPUs and delivers the return on investment needed to sustain the next wave of AI innovation. We've just delivered the fastest storage on the planet, that is purpose built for the AI Era."

Keystone Sovereign targets European compliance

Keystone, NetApp's storage subscription service, gains a Sovereign variant aimed at European customers whose sovereignty requirements now extend beyond data residency to operational access, telemetry, and data flow transparency. New capabilities include European-based data residency for customer data, logs, backups and telemetry; European-based support and escalation paths; European-controlled access and management processes; documented telemetry and data flows; and contractual and legal structures reinforcing sovereignty expectations.

Ivan Iannaccone, VP and GM, Keystone at NetApp, said: "With NetApp Keystone Sovereign, we are helping qualified European customers address these evolving expectations with additional regional, support, access, and data-handling controls designed to build trust and keep critical data and support within defined geographic boundaries."

Initial pilots will cover Germany and France, with additional countries planned. The roadmap implications for other territories are clear.

Fleet management and ChatOps

NetApp Console adds a second delivery model: Console Fleet Management installs directly in the customer environment, letting admins manage an entire estate as a single entity with fleet management, observability, and AI. Console Autonomous Operations provide AI-assisted operations with human oversight, using agentic AI-driven orchestration to predict issues, resolve incidents faster, and optimize capacity and costs. A conversational ChatOps interface replaces manual workflows with intent-based provisioning, policy guardrails, AI-powered reporting, predictive analytics, and automated remediation across ONTAP systems. It operates through an open LLM gateway, so customers can use their own models, with all actions kept within defined storage classes, policies, and governance boundaries.

Sandeep Singh, NetApp SVP and GM, Platform, said: "Customers should not have to choose between the performance and control of on-premises infrastructure and the flexibility of cloud. With these new hybrid multi-cloud capabilities, NetApp is helping enterprises break down the borders between environments, operate with greater autonomy, and put efficiency, sovereignty, and control at the center of their data strategy."

AI Data Engine expands metadata discovery

The AI Data Engine (AIDE), NetApp's AI-focused ETL pipeline built on the AFX metadata engine, now discovers and characterizes data across the full enterprise estate — including ONTAP, StorageGRID, and non-NetApp storage over NFS, SMB and S3. NetApp says this helps customers reduce security risk, optimize storage, accelerate migrations, and identify AI-ready data in place. Previewed additions include deeper governance intelligence, open analytics integrations with Starburst and Onehouse.AI, and new agentic services.

Partnership wave: Oracle, SAP, Commvault, Supermicro, Nutanix

The largest cloud announcement is a fully managed, OCI-native NetApp storage service bringing ONTAP to Oracle Cloud Infrastructure, paralleling existing native ONTAP implementations on AWS, Azure and GCP. General availability is planned within the next 12 months. Konstantinos (kp) Papamiltiadis, SVP of Partnerships at Oracle Cloud Infrastructure, said: "Enterprises shouldn't have to rebuild their data infrastructure to move to the cloud. OCI NetApp Storage Service gives customers a trusted, familiar path to OCI, so they can focus on what the data does, not how it's managed."

NetApp and SAP intend to explore deeper cloud-native integration of NetApp capabilities into SAP Cloud Infrastructure, building on existing file and block storage services, and will evaluate StorageGRID object storage there, including a technical validation. Commvault is partnering with NetApp to protect AI workload data, connecting NetApp's autonomous ransomware protection signals on primary storage to Commvault recovery workflows so joint customers can "recover in minutes rather than days or weeks without introducing reinfection." Pranay Ahlawat, Chief Technology and AI Officer at Commvault, said: "By integrating NetApp's ransomware detection and immutable data protection with Commvault's clean recovery capabilities, we help customers accelerate response, restore trusted data with confidence, and reduce mean time to clean recover (MTCR)."

The Supermicro collaboration is the foundation under Novus: jointly validated AI infrastructure spanning AI factories, enterprise deployments, neoclouds, and sovereign AI initiatives, with Novus Data Director on qualified Supermicro servers and AFF A90 as the ONTAP data layer. The companies will also deliver NetApp AIPod with Supermicro, a converged architecture built to the Nvidia Enterprise Reference Architecture, aimed at AI training, inference, and agentic workloads from dozens to hundreds of GPUs. Alvaro Celis, NetApp SVP, Chief Commercial, Partner and Ecosystem Officer, said: "By combining Supermicro's AI leadership and ability to scale with the NetApp Platform, we can give customers—from commercial, enterprise to internet-with the most demanding requirements—secure, reliable AI solutions they can deploy quickly to accelerate innovation and build market leadership." Separately, ONTAP now combines with Nutanix Cloud Infrastructure so VM and containerized workloads can use ONTAP storage features.

NetApp describes Novus explicitly as an initial release, signaling more capacity and capability tiers to come as it fights DDN, WEKA, VAST Data and Everpure for the AI factory storage market rather than conceding ground to AI-focused challengers.