Pranav Hotkar
Mumbai, India - March 4, 2026 - Aqylon Nexus Limited has signed a memorandum of understanding (MoU) with DataDirect Networks to explore collaboration on AI-ready hyperscale data center infrastructure designed to support high-performance computing and generative AI workloads.
The agreement, disclosed through a regulatory filing, outlines a framework for the two companies to jointly evaluate opportunities to develop large-scale data center facilities optimized for artificial intelligence and data-intensive applications. The proposed collaboration aims to support emerging workloads such as large language model training, generative AI, and enterprise high-performance computing.
Under the MoU, Aqylon Nexus will explore opportunities to develop hyperscale facilities, while DataDirect Networks will provide expertise in AI-optimized storage and data platform technologies. The partnership will evaluate the deployment of high-performance storage architectures designed to handle large volumes of AI training and inference data.
DataDirect Networks is widely known for its high-throughput storage platforms used in supercomputing and AI research environments. Its technologies are commonly deployed in high-performance computing clusters where extremely fast data access and parallel processing are required to support advanced analytics and machine learning applications.
Aqylon Nexus, which has been repositioning its strategy toward digital infrastructure and AI ecosystem development, said the collaboration aligns with its broader vision of building next-generation infrastructure capable of supporting rapidly growing AI computing demand.
The companies said the partnership will explore potential projects in India as well as other mutually agreed international markets. The collaboration may include evaluating advanced data architectures that combine NVMe-based storage, parallel file systems, and AI-optimized data pipelines designed for large-scale computing environments.
According to the regulatory disclosure, the memorandum is non-binding and establishes a preliminary framework for cooperation between the two companies. The agreement will remain valid for one year unless extended by mutual consent.
Either party may terminate the MoU with prior written notice during the term of the agreement.
If implemented, the collaboration would focus on designing data center infrastructure capable of supporting the growing requirements of AI training and inference workloads, which increasingly demand high-density compute clusters and ultra-fast data movement across storage and processing systems.
The announcement reflects a broader industry shift toward specialized infrastructure tailored for AI computing, where storage performance and data throughput play a critical role in enabling large-scale model development and deployment.
