The platform functions as an AI-native infrastructure, designed to transform proprietary documents and unstructured data into traceable, searchable knowledge. By integrating document processing, vector retrieval, and AI-service management into a single system, Seahorse allows organizations to maintain data control while deploying RAG services. Users can cross-reference AI-generated outputs with original source documents, a feature critical for enterprise verification.
At the FMS 2026 conference, Dnotitia also unveiled its VDPU, a custom processor built specifically for vector search and graph traversal. By moving vector operations closer to the stored data, the chip aims to offload host CPU tasks and minimize data movement. The company is currently showcasing this technology via a new vector-search accelerator card at Booth 1144, seeking partnerships with global storage vendors for joint commercialization and proof-of-concept projects.
CEO MK Chung noted that AI bottlenecks are increasingly centered on storage and retrieval rather than just raw computing power. Beyond the winning application, Dnotitia was recognized as a finalist for the Startup Business Growth Award, reflecting its ongoing development in semiconductor design and global enterprise engagement. The company will continue its product demonstrations through August 6.
Comments (0)
No comments yet. Be the first!