KIOXIA, a leader in flash memory solutions, announced an update to its AiSAQ™ (All-in-Storage ANNS with Product Quantization) software on July 3, 2025, enhancing retrieval-augmented generation (RAG) systems. The open-source release introduces configurable controls that allow system architects to balance search performance and vector capacity, optimizing solid-state drive (SSD) usage without hardware changes. This advancement supports scalable AI applications by reducing dependence on costly DRAM.
KIOXIA AiSAQ update released as open-source on July 3, 2025.
Enables flexible tuning of search performance vs. vector capacity in RAG systems.
Optimized for SSDs, eliminating DRAM dependency for vector searches.
Supports billion-scale datasets with minimal memory usage.
Ideal for RAG and vector-intensive applications like offline semantic searches.
First introduced in January 2025, KIOXIA AiSAQ uses a novel approximate nearest neighbor search (ANNS) algorithm tailored for SSDs, enabling vector database searches without storing index data in DRAM. The latest update introduces flexible configuration options, allowing architects to adjust the trade-off between search performance (queries per second) and vector capacity within fixed SSD storage. Higher performance consumes more SSD capacity per vector, reducing total vectors, while maximizing vectors lowers performance. This update ensures optimal balance for diverse workloads, enhancing scalability for RAG and other vector-heavy applications like offline semantic searches.
Traditional ANNS algorithms rely on DRAM for high-speed performance, but DRAM’s limited capacity and high cost constrain large-scale AI systems. KIOXIA AiSAQ enables direct SSD-based searches, supporting billion-scale datasets with negligible memory usage and fast index switching. This approach eliminates the need to load index data into DRAM, enabling instant database launches and seamless switching between user- or application-specific databases on the same server, ideal for cloud systems with disaggregated storage.
“With the latest version of KIOXIA AiSAQ software, we’re giving developers and system architects the tools to fine-tune both performance and capacity,” said Neville Ichhaporia, senior vice president and general manager of the SSD business unit at KIOXIA America, Inc. “This level of flexibility is critical to building scalable, RAG systems – powered by SSD storage.” By open-sourcing AiSAQ, KIOXIA fosters broader adoption of SSD-centric AI architectures, making high-performance RAG systems more accessible.
The update complements KIOXIA’s high-capacity LC9 Series NVMe SSDs (122.88 TB), designed for AI applications with 8th-generation BiCS FLASH™ 3D technology. These SSDs enhance AiSAQ’s ability to handle large-scale datasets, offering a cost-effective alternative to DRAM for AI training, inference, and fine-tuning.
KIOXIA’s AiSAQ update positions it as a game-changer in AI infrastructure, enabling efficient, scalable RAG systems. By leveraging SSDs and reducing DRAM reliance, KIOXIA empowers developers to build cost-effective, high-performance AI solutions, driving innovation across industries.
KIOXIA America, Inc. is the U.S.-based subsidiary of KIOXIA Corporation, a leading worldwide supplier of flash memory and solid-state drives (SSDs). From the invention of flash memory to today’s breakthrough BiCS FLASH™ 3D technology, KIOXIA continues to pioneer innovative memory, SSD and software solutions that enrich people's lives and expand society's horizons. The company's innovative 3D flash memory technology, BiCS FLASH, is shaping the future of storage in high-density applications, including advanced smartphones, PCs, automotive systems, data centers and generative AI systems.
TechIntelPro empowers B2B technology decision-makers with the latest industry trends, deep insights, leadership perspectives and actionable intelligence to drive both organizational and professional milestones—and thrive on the tech pulse.