NetApp acquires DataPelago to bring GPU acceleration to the storage layer. The deal integrates DataPelago's Nucleus engine, which processes data at the storage layer using CPUs and GPUs, cutting infrastructure costs by up to 80% and delivering 10x faster performance than conventional approaches. NetApp faces growing competition from Pure Storage and Dell as enterprises race to put AI into production.
NetApp Inc. acquired DataPelago, a startup whose Nucleus engine processes data at the storage layer using CPUs and GPUs, cutting infrastructure costs by as much as 80% and eliminating the biggest bottleneck in enterprise AI deployment.
"The key to accomplishing this objective is to enable accelerated computing where the data is created and stored," George Kurian, chief executive officer at NetApp, said.
DataPelago's Nucleus engine delivers performance up to 10 times faster than conventional approaches by processing data where it lives rather than moving it to external compute clusters, the companies said. The technology eliminates the need to copy data from operational systems to AI systems — a process known as zero-copy activation. NetApp manages more enterprise data across more environments than any other company in the industry, according to Syam Nair, its chief product officer.
NetApp, which trades at roughly 18 times forward earnings, faces competition from Pure Storage Inc. and Dell Technologies Inc. in the enterprise storage market. The DataPelago deal signals NetApp's push to differentiate through AI-native infrastructure, following recent partnerships with Cisco Systems Inc., Google Cloud, Red Hat Inc. and SK Telecom.
Why processing at the storage layer matters
Enterprise AI deployment today requires copying data from operational databases to separate GPU clusters for training and inference — a process that can take weeks and consumes significant network bandwidth. DataPelago's Nucleus engine uses heterogeneous accelerated computing across CPUs and GPUs to process data directly at the storage layer, bypassing the data movement bottleneck entirely.
The approach addresses a growing pain point: enterprises have invested billions of dollars in GPUs and AI models, but their data remains fragmented across silos, leaving computing resources idle. "Enterprises have invested billions in GPUs and AI models, but their data remains fragmented, leaving valuable computing resources to sit idle rather than putting these investments to work," Rajan Goyal, founder and chief executive officer of DataPelago, said.
Who wins, who loses
The acquisition strengthens NetApp's position against Pure Storage, which has been pushing its own AI-ready storage offerings, and Dell, which dominates enterprise storage through its PowerScale line. Both competitors have been adding GPU-direct capabilities to their storage platforms.
For hyperscaler cloud providers — Amazon Web Services Inc., Microsoft Corp.'s Azure and Google Cloud — the deal is less immediately threatening. NetApp's ONTAP storage software is already natively embedded in all three major clouds, meaning DataPelago's technology could eventually be offered as a cloud service rather than competing with cloud-native storage.
DataPelago will operate as a wholly owned subsidiary of NetApp following the acquisition. The companies did not disclose the deal's financial terms.
NetApp shares, which have gained about 12% this year, trade at a discount to Pure Storage's 28 times forward earnings. The acquisition could help close that valuation gap if DataPelago's technology drives incremental storage revenue from AI workloads, though the timeline for integration remains unclear.
This article is for informational purposes only and does not constitute investment advice.
