NetApp has acquired DataPelago, a California startup whose software runs data processing across CPUs and GPUs at the storage layer, adding a missing piece to the storage vendor’s pitch that enterprise data infrastructure should do more than store and protect information. Terms were not disclosed.
DataPelago’s core product, Nucleus, is designed to process data where it already resides rather than shuttling it into separate analytics or AI clusters. That matters because the expensive part of many enterprise AI projects is no longer the model. It is the plumbing. Companies have spent heavily on GPU capacity, but much of that compute sits underused while data is copied, reformatted, governed, and moved across systems before it can feed training or inference workloads.
NetApp is buying a way to move closer to the compute budget. For years, storage vendors have argued that data gravity would make the infrastructure layer central to AI. DataPelago gives NetApp a more concrete answer to that claim by embedding accelerated processing into the point where enterprise data is created, managed, and retained. If the technology integrates cleanly with ONTAP and NetApp’s cloud partnerships, the company can sell not just storage capacity but faster AI readiness and lower infrastructure overhead.
The acquisition also reflects a broader shift in enterprise infrastructure buying. Customers are increasingly less interested in assembling separate storage, ETL, governance, and AI orchestration stacks if a platform vendor can collapse some of that complexity. That creates an opening for incumbents with installed bases and distribution, while narrowing the room for standalone data pipeline startups.
Acquire.fyi data shows technology M&A volume is down 12.6% year to date, even as total deal value in the sector has climbed 73.5%, a sign that buyers are concentrating capital behind assets tied to AI architecture and platform control rather than broad software rollups. NetApp’s move fits that pattern. It is a targeted infrastructure bet aimed at defending relevance as AI spending shifts from experimentation to production. The next question is whether rivals in storage and data management respond with similar tuck-ins, or risk becoming passive repositories in an increasingly active data stack.
Source: Company press release and Acquire.fyi's proprietary data
