Elasticsearch
Vector databases are eating its lunch.
Elasticsearch is a powerful full-text search engine that has served well for years. But the AI wave has shifted the search paradigm toward semantic similarity, and purpose-built vector databases handle that workload better.
Qdrant, Pinecone, and Weaviate are optimised for vector operations from the ground up. Elasticsearch bolted on vector search as an afterthought. For hybrid search combining keywords and semantics, the vector-native tools now offer full-text search too — making Elasticsearch's position increasingly narrow.
If you already run Elasticsearch for log aggregation or traditional search, keep it. For new search workloads, especially anything involving embeddings or AI, start with a vector database.