PuppyGraph Raises $10M Series A to Let AI Agents Query Enterprise Data as a Graph
PuppyGraph raised a $10 million Series A for its graph query engine, which treats warehouse and lakehouse data as a single graph without moving it, adding new tooling for AI agents.
PuppyGraph announced a $10 million Series A on November 10, bringing its total funding to $15 million, following the general availability of version 1.0. The company offers a graph query engine that treats data already stored in warehouses and lakehouses, including Iceberg, Databricks, Snowflake, ClickHouse, PostgreSQL, DuckDB, Hudi, and Fabric, as a single graph without moving the data.
The company said AI agents can work across 3,000 tables with query results returned in under one second, and that a 10-hop query across billions of edges completed in 2.26 seconds on a four-machine cluster; those figures are vendor-reported. New features include an AI-assisted schema builder that automatically maps tables into entities and relations, though developers must review and approve a schema before agents can use it. Queries referencing undefined relations are rejected with actionable errors. The release adds Model Context Protocol (MCP) integration and short-lived credentials tied to user identity, with row-level permissions enforced per hop.
Deployment runs from a single Docker command, and a free version is available without sign-up.