NetApp unveils zero-copy data activation for agentic AI
NetApp (NASDAQ: NTAP) has announced a set of platform updates designed to help enterprises deploy agentic AI workloads against their existing data estates without replicating or moving that data first. The centrepiece is an expanded version of the company's AI Data Engine (AIDE), which now supports heterogeneous metadata discovery across NetApp ONTAP, StorageGRID, and third-party storage systems via NFS, SMB, and S3 interfaces. A deepened integration with Commvault extends that foundation to cover ransomware recovery, connecting NetApp's autonomous threat-detection signals on primary storage directly to Commvault's orchestrated clean-recovery workflows.
The announcements were made at an event in Dubai and follow a broader industry shift in which AI agents, rather than batch analytics pipelines, are becoming the primary consumers of enterprise data. That shift creates a structural problem: most enterprise data estates were built for human-paced access, fragmented across file, object, cloud and on-premises environments, and typically prepared for AI through copy-heavy extract-transform-load pipelines that introduce latency, cost and compliance risk.
What the platform updates deliver
AIDE's expanded metadata capability is positioned by NetApp as the mechanism for discovering and classifying data across a mixed storage estate without first centralising it. The company says this allows customers to identify AI-ready datasets, reduce security exposure, and optimise storage utilisation as an immediate outcome, before moving on to richer capabilities such as semantic search and knowledge graphs, which remain in preview. Forthcoming integrations with analytics engines Starburst and Onehouse.AI were also previewed, though NetApp noted that statements about unreleased features carry no commitment on timing or availability.
The Commvault joint solution adds near-real-time threat detection, isolated cleanroom validation, and what the two companies describe as recovery measured in minutes rather than days. Pranay Ahlawat, Chief Technology and AI Officer at Commvault, said that tighter integration of NetApp's immutable data protection with Commvault's clean-recovery stack helps customers "reduce mean time to clean recover" as AI-driven attacks compress the window available for response.
Shari Lava, GVP of AI Research at IDC, described NetApp's approach as "a useful example" of connecting data discovery, governance, and cyber resilience at the storage layer, arguing that vendors capable of doing so give customers "a more pragmatic path to prepare data for AI, without the risk, cost, and complexity that often come with large-scale data movement."
Market context and competitive dynamics
The zero-copy or in-place activation framing puts NetApp in direct competition with data lakehouse vendors such as Databricks and Snowflake, which have built large ecosystems around centralised data platforms, as well as with pure-play AI data pipeline vendors. Cloud hyperscalers are pursuing a parallel path through native storage and governance services tied to their own AI platforms, creating a lock-in risk that NetApp explicitly positions its disaggregated architecture against.
Ransomware resilience for AI workloads is an emerging competitive differentiator. As enterprises commit sensitive operational data to agentic systems, the blast radius of a successful attack expands, and the tolerance for extended recovery windows shrinks. The NetApp-Commvault combination targets this concern directly, at a moment when regulators under DORA in Europe and NIST's Cybersecurity Framework in the US are raising expectations for documented recovery time objectives across critical infrastructure operators.
NetApp did not disclose customer names, adoption figures, or commercial terms for the updated platform. Near-term milestones to watch include the general availability date for the previewed AIDE governance and agentic-activation features, and any named enterprise reference customers for the Commvault joint solution.