US Equities insigtX

Dell Bets on Liquid Cooling and Software-Defined Storage to Support AI Factory Data Throughput

The density of AI factories has become so high that storage—not just compute—now determines how quickly they turn data into tokens. For Dell Technologies, the answer begins with…

Published
Market
US Equities
Source
insigtX

The density of AI factories has become so high that storage—not just compute—now determines how quickly they turn data into tokens. For Dell Technologies, the answer begins with transforming its PowerEdge servers into flexible, software-defined storage nodes that can change roles as workloads evolve. The shift from training to inference and agents is putting new pressure on time-to-value. George Papadopoulos, Dell's senior vice president of products and engineering, said Dell has been expanding its AI data platform this year, adding a semantic layer for agentic workloads, and its vision of the AI factory now extends far beyond compute and graphics processing units.

"We look at the AI factory not just as a set of CPUs and GPUs—networking is becoming increasingly important, especially in terms of speed, reliability of storage, and the ability to generate context and knowledge from raw data," Papadopoulos said. Papadopoulos spoke with John Furrier and Rebecca Knight of theCUBE Research during an exclusive broadcast on theCUBE, SiliconANGLE Media's live studio, at the SC26 conference.

PowerEdge storage built to adapt to shifting AI workloads Customers want to buy storage that meets current needs and can be repurposed later. Papadopoulos explained that, for example, a customer might start with file storage and, as workloads evolve, move to object storage or parallel file storage. Agentic workloads rely on small input/output operations rather than streaming, so the right storage type can be switched quickly. Dell's answer is Dell Exascale Storage, which runs as licensed software on Dell servers.

"This is a software-defined solution we truly built for a specific purpose, allowing us to take PowerEdge-based nodes... through flexible licensing, enabling you to deploy file roles, block roles, object roles, or parallel file roles, and allowing that flexibility to change over time as your needs evolve," Papadopoulos said. Density is also forcing a rethink of cooling. According to Papadopoulos, the power draw of today's AI systems is so broad that customers are planning data centers with 100% liquid cooling. That requirement has now extended to storage.

"So the discussion really goes deep into how we most effectively deliver direct liquid cooling solutions, and on the storage side, that even means liquid cooling the drives themselves, which is something that hasn't been done to date," Papadopoulos said. Dell used SC26 to launch a 2U, 32-drive node with 100% liquid cooling, touting it as the first direct liquid-cooled storage for rack-scale AI. The node deploys as an Exascale system and can play any role. Separating software from the appliance also allows Dell to bring new generations of PowerEdge to market faster.

Papadopoulos noted that this flexibility is critical for new cloud operators, who need an exit path for every purchase. "It allows them to make upfront purchases and know they won't be locked into a particular storage type," Papadopoulos said, "and they won't end up with a set of data silos that can't be leveraged in future development." Below is the full video interview, part of SiliconANGLE and theCUBE's coverage of SC26:

(* Disclosure: Dell Technologies sponsored this segment on theCUBE. Dell Technologies and other sponsors have no editorial control over content on theCUBE or SiliconANGLE.)

Original: https://siliconangle.com/2026/11/17/poweredge-nodes-turn-software-defined-storage-ai-dellatsc26/

insigtX content is informational and educational, not investment advice.