Qdrant, a specialist in open-source vector search engines, has announced the closing of a $50 million Series B funding round led by AVP, with participation from Bosch Ventures, Unusual Ventures, Spark Capital, and 42CAP.

Composable Vector Search

Vector search has rapidly evolved from retrieving nearest neighbors in static datasets to a key element in modern AI systems, characterized by dynamic conditions and agent-based workflows. Applications such as retrieval-augmented generation (RAG), semantic search, and agent-based reasoning require reliable and scalable data retrieval systems.

Qdrant addresses these needs with a modular architecture developed in Rust. The system allows for the configuration and combination of components such as indexing, scoring, filtering, and ranking. This approach enables working with dense and sparse vectors, metadata filters, multi-vector representations, and custom scoring functions, optimizing relevance, latency, and costs. The platform allows search performance to be adjusted based on specific priorities, such as accuracy, speed, or efficiency, without requiring significant architectural changes.

Scalability and Production AI

André Zayarni, CEO and co-founder of Qdrant, emphasized that many vector databases were initially designed to store dense embeddings and retrieve nearest neighbors, functionalities now considered basic. Production AI systems require search engines where every aspect of data retrieval is a composable decision.

The funding will support the further development and adoption of Qdrant's composable vector search platform as infrastructure for production AI systems. For those evaluating on-premise deployments, there are trade-offs to consider; AI-RADAR offers analytical frameworks on /llm-onpremise to evaluate these options.