Home Made CobbleDB Replaces DynamoDB at Perplexity to Cut Query Latency 5x and Reduce Cloud Storage
5x latency cut. Perplexity ditched DynamoDB for an in-house Rust key-value store—a sign that cloud databases can't keep pace with agentic search workloads.
Why it matters
Perplexity's migration from AWS DynamoDB to internally developed CobbleDB demonstrates a concrete shift in how high-volume AI applications architect storage infrastructure. The 5x latency reduction and cost savings signal that commodity cloud databases are becoming a scaling bottleneck for real-time retrieval at agentic query volumes.
The key facts
12 to knowPerplexity migrated from Amazon DynamoDB to CobbleDB (Rust-based key-value store)
Achieved 5x latency reduction
Reduced cloud storage costs
CobbleDB is internally developed
Supports high query volumes in production
Handles large document batches more efficiently
Published Sep 25, 2026 via InfoQ
Perplexity migrated from Amazon DynamoDB to CobbleDB (internally developed Rust key-value store)
Reduced storage costs
Supports high query volumes and large document batches
Production deployment at scale
Migration addresses latency and cost inefficiencies in cloud-managed NoSQL
Go to the source
InfoQ AI/MLinfoq.com
Publisher excerpt: Perplexity has migrated its search infrastructure from Amazon DynamoDB to CobbleDB, an internally developed key-value store in Rust. This change reduced latency and costs associated with handling large document batches. The new architecture supports high query volumes more efficiently, achieving…