NetApp acquires DataPelago to bring GPU-accelerated processing to storage layer
The AMW Read
Acquisition adds a new entrant to the data infrastructure player map (05.§2 update); segment-level significance as it signals storage-compute convergence.
NetApp acquires DataPelago to bring GPU-accelerated processing to storage layer
NetApp (NTAP) has acquired DataPelago, a startup specializing in AI data infrastructure. The deal brings DataPelago's Nucleus universal data processing engine into NetApp's portfolio, enabling GPU-accelerated data processing directly within the storage layer. This "zero-copy activation" approach allows enterprises to prepare, govern, and transform datasets without moving them to separate AI infrastructure, potentially delivering up to 10x faster processing and 80% infrastructure cost savings.
Why it matters: This acquisition exemplifies the acqui-licensing pattern in data infrastructure, where an incumbent storage player acquires a novel data-processing capability rather than building it internally. It also signals the growing convergence of storage and compute in enterprise AI — a structural force (cross.§A Compute Economics) as GPU scarcity pushes optimization to the data layer. NetApp is betting that DataPelago's heterogeneous computing engine can differentiate it against HPE, Teradata, and cloud providers in the competitive AI infrastructure market.
Expert take: NetApp's move mirrors HPE's acquisition strategy with Determined AI and Pachyderm — buying into the AI stack rather than organically developing. The 80% cost reduction claim, if credible, would materially improve ROI for enterprises running GPU clusters, but integration risk and competitive responses from cloud-native architectures remain open questions. The deal does not disclose financial terms, so capital-cycle dynamics are unclear.