NetApp has acquired DataPelago, a startup focused on GPU-powered data processing, to tackle what it identifies as the next critical bottleneck in AI storage infrastructure.
DataPelago’s technology enhances AI workloads
The acquisition brings DataPelago’s universal data processing engine into NetApp’s offerings. This allows customers to execute AI workloads directly on stored data without first transferring it. The technology works across various hardware types—CPUs, GPUs, or future accelerators—and supports any query engine, as stated in DataPelago’s public materials.
DataPelago’s main product, DataPelago Accelerator for Spark, claims to reduce costs by up to 80% for eligible workloads while improving performance. The tool eliminates the need for application rewrites or data migration, which appeals to enterprises managing large datasets.
The purchase price remains undisclosed. DataPelago, which launched publicly in October 2024, had secured $47 million from investors including Eclipse, Taiwania Capital, Qualcomm Ventures, Alter Venture Partners, Nautilus Venture Partners, and Silicon Valley Bank, a division of First Citizens Bank.
Related: 7AI Launches First Formal Channel Program To Drive Agentic SOC
NetApp CEO George Kurian described the acquisition as a solution to the widening gap between AI model capabilities and the infrastructure required to support them. “As AI models and the chips that power them get ever more effective, enterprises need data infrastructure that is just as intelligent and powerful to harness the potential of their data,” he said. Kurian added that the move would help customers process data with the agility required to unleash competitive advantage.
For most organizations, the primary obstacle isn’t developing AI models but ensuring those models can access the right data efficiently. Older storage systems weren’t built for AI’s demands, where latency and throughput determine success. DataPelago’s method, which places processing power closer to the data, may reduce delays that currently hinder AI projects.
The transition to AI-driven workflows requires storage systems to serve two distinct users: people and machines. People might search for a specific file, while an AI model may need to analyze millions of records in seconds. Conventional storage systems struggle to optimize for both use cases.
Partner perspective on the acquisition
Dave Van Hoy, founder and president of Advanced Systems Group, an Emeryville, Calif.-based solution provider and NetApp channel partner, told CRN that prior to the acquisition, he had not heard of DataPelago but is looking forward to seeing it become a part of NetApp. “I’m super excited about it because it aligns to my view of where storage is going and what’s important around that,” Van Hoy said.
Van Hoy, who also collaborates with IBM, noted the issue isn’t new but has grown more pressing. In media and entertainment, companies depend on media asset management systems, which often require extensive manual input. Outside those fields, employees reportedly spend up to one-fifth of their time locating data they know exists somewhere in their systems.
Related: Timeless Beauty: Vintage-Inspired Engagement Rings in Londonderry
“That view really comes from some work I’ve been doing with IBM, who I think, ironically, is way out in front of this,” Van Hoy said. He explained that IBM believes in the concept of creating content-aware storage that needs to be accessed by both human end users and machines focused on AI. In both cases, he said, the problem is the same: how to find needed data as fast as possible with the least amount of friction.
NetApp provided no further details beyond its press release, stating more information would come later. The company plans to report its fiscal first-quarter 2027 earnings on September 2.
The acquisition reflects a wider industry effort to redesign storage for AI. As models expand and datasets grow, the traditional division between storage and computing creates inefficiencies. NetApp’s strategy involves bringing processing closer to data, helping customers avoid the expense and time of moving data to external compute clusters.
Integration remains the priority. NetApp hasn’t announced when DataPelago’s technology will be fully available in its products, but the emphasis on compatibility—no rewrites or migration—suggests a straightforward path for current users.
