How HubSpot Scaled Semantic Search to 20 Billion Vectors
20 billion vectors. That's how HubSpot scaled semantic search from POC to production—now powering agents and RAG across 38+ teams.
Why it matters
HubSpot's infrastructure scaling story reveals the operational challenges SaaS companies face when vector databases move from experimental to mission-critical. As agent adoption accelerates, retrieval latency and vector management become competitive advantages, not nice-to-haves.
The key facts
9 to know20 billion vectors managed in production
38+ internal teams using the platform
System now supports agents, RAG, and contact deduplication
Increased agent usage driving retrieval quality and latency requirements
Evolved from POC to internal service infrastructure
20 billion vectors managed across 38+ teams
Semantic search evolved from POC to internal service layer
Agent usage growth has elevated retrieval quality and latency as primary concerns
Published July 2026 — signals enterprise AI infrastructure maturation
Go to the source
InfoQ AI/MLinfoq.com
Publisher excerpt: SaaS software vendor HubSpot has described how its semantic search platform grew from a proof of concept into an internal service that now manages more than 20 billion vectors across 38-plus teams. The company says the system now supports agents, RAG, and contact deduplication, and that the…