Data intelligence and storage company DDN is reportedly eyeing further investment to more accurately represent the value it delivers to its clients.

Bloomberg reports the Chatsworth, California-based vendor, formerly DataDirect Networks, is looking at a potential investment round toward the end of the year, a move CEO Alex Bouzari said was aligned with “bringing more savvy, sophisticated, smart people into our circle.”

DDN offers high-performance data storage and management infrastructure designed to accelerate AI workloads. Among its offerings is the software-defined data platform Infinia 2.0, which unifies AI datasets across on-premise, edge, and cloud environments, and xFusionAI, which connects its high-speed parallel file system with elastic object storage onto a single hardware platform that lets users train large models and run real-time inference using the same data pool.

The firm already secured some $300 million from Blackstone in January, with the asset management giants valuing DDN at $5 billion. That came off the back of industry reports that the firm was mulling a potential IPO.

DDN was founded in the late 90s, but in recent years, it’s firmly focused on supporting AI infrastructure efforts. According to Bloomberg, it counts clients including xAI and the neocloud Lambda.

Earlier this month, it revamped its AI data intelligence platform to support agentic workloads. Chief of which saw the firm align with Nvidia’s BlueField-4 STX reference architecture to help enterprise customers scale secure AI environments for training and inference. Powered by Nvidia accelerated computing, DDN's platform lets organizations operationalize secure AI environments by combining high-performance data orchestration and multi-tenant isolation with real-time services optimized for training, inference, vector databases, retrieval-augmented generation (RAG) pipelines, and autonomous workflows.

DDN’s agentic updates followed additions to its Lustre platform that allow users to share key-value (KV) cache to boost AI inference workloads.

The offering, which DDN manages alongside Google Cloud, employs a shared cache layer across inference clusters rather than keeping KV-cache in each server's local memory. As a result, DDN and its hyperscaler helper claimed total inference throughput was improved by as much as 75%.