The startup is developing a graph query engine that allows AI agents to investigate connected information without copying it into a separate database.
PuppyGraph has raised $10 million in a Series A led by Particle Future, with participation from defy.vc, bringing total funding to $15 million.
Led by co-founder and CEO Weimo Liu, the company develops graph query infrastructure designed to work with existing enterprise data platforms.
Its technology enables organizations to analyze relationships across datasets without having to move or duplicate the underlying information.
Why Traditional Data Queries Are Not Enough
Enterprise information is often distributed across thousands of tables.
A company may store customers, suppliers, products, transactions and operational activity in different systems.
Traditional SQL queries can answer well-defined questions, but complex investigations may require following relationships across many datasets.
Graph queries make it possible to explore those connections.
PuppyGraph provides a graph query layer that operates over existing data, allowing users to investigate relationships between entities.
For example, an AI agent investigating falling sales could connect delayed orders to particular products, components and suppliers.
The discovery might come from several steps of investigation rather than one predefined query.
Infrastructure for Autonomous Analysis
As AI agents become capable of planning and executing their own investigations, they need access to connected enterprise information.
PuppyGraph wants to provide that layer.
Its architecture is designed to query information where it already exists, avoiding the cost and complexity of maintaining a separate graph database containing duplicated records.
The approach could help organizations preserve existing investments in data warehouses and analytical platforms.
The Next Challenge for Enterprise AI
PuppyGraph competes with graph database vendors and emerging AI data infrastructure providers.
Its opportunity depends on whether autonomous agents will require a specialized relationship-query layer as they become more sophisticated.
The Series A will support product development and broader enterprise adoption.
Its next challenge is demonstrating that graph-based investigations consistently produce useful business insights while maintaining performance and security at scale.
Sources: PuppyGraph announcement; Axios (10 November).