Kioxia Brings Cost-Efficient, Scalable Vector Search to the Milvus Ecosystem

Kioxia Corporation announced that its Nearest Neighbor Search (ANNS) software technology, KIOXIA AiSAQ™, has been officially integrated into the open-source vector database starting with Milvus version 2.6.4. This integration allows Milvus users to fully leverage KIOXIA AiSAQ™'s SSD-optimized vector search capabilities, providing developers and enterprises with a practical and cost-effective path to scaling AI applications without the DRAM memory expansion challenges typically associated with large-scale vector searches.

The AI industry is shifting from building large-scale foundational models to deploying scalable, cost-effective inference solutions to address real-world challenges. Retrieval Augmentation (RAG) is central to this transformation, and KIOXIA AiSAQ™ technology was developed to assist the community in leveraging SSD-based vector architectures. Its integration into the Milvus ecosystem not only lowers the adoption barrier for the open-source community but also supports developers in building faster and more efficient AI applications.

First announced earlier this year, KIOXIA AiSAQ™ is an open-source software technology that significantly improves vector scalability *1 by storing all RAG-related database elements on an SSD . With DRAM scalability becoming a critical bottleneck for massive inference and RAG workloads, KIOXIA AiSAQ™ technology represents a breakthrough, significantly reducing DRAM requirements while maintaining high-quality vector search accuracy.

KIOXIA AiSAQ™ technology is now integrated into Milvus, and Kioxia and the open-source community are building a new category of scalable, cost-effective vector search solutions designed to meet the rapidly growing needs of modern AI applications