DDN launches xFusionAI, optimizing AI model performance

DDN has launched xFusionAI, an Artificial Intelligence (AI) infrastructure that consolidates both training and inference capabilities into one optimized platform. This solution is tailored for enterprises, hyperscalers, and research institutions, aiming to enhance the performance and efficiency of AI workflows.

xFusionAI integrates DDN’s EXAScaler parallel file system with Infinia’s AI-native scalability and performance, creating a hybrid platform designed to handle a wide range of AI models, from lightweight architectures to those with trillions of parameters.

Key features include a 100x acceleration in Large Language Model (LLM) training, a 10x increase in real-time performance, and a 60% reduction in inference costs. Early adopters have reported substantial improvements in AI workflows.

“xFusionAI combines the raw performance of EXAScaler with the intelligent scalability of Infinia,” said Sven Oehme, CTO at DDN, emphasizing its role in supporting diverse AI applications.

This new solution enables seamless scaling across various environments, promoting efficient AI deployment in on-premises (on-prem), cloud, and air-gapped systems. DDN focuses on providing an end-to-end AI infrastructure that supports faster training and more intelligent inference for diverse business applications.