Partner Alliance Director, Neo4j
7 min read
Modern data warehouse providers like Snowflake deliver a consolidated platform for an organization’s data and AI ecosystem. Large Language Models (LLMs) offer extraordinary evolutionary capabilities in human-to-IT interactions. Yet traditional analytics and vector searches alone often fail to uncover the deep, complex relationships hidden across enterprise data. This slows down decision-making and limits the value of existing data assets. It can also cause critical errors for generative AI applications. To solve this, Neo4j and Snowflake have partnered to bring enterprise context to Snowflake’s AI Data Cloud. This partnership enables organizations to gain deeper insights, drive innovation, and set a new standard for intelligent decision-making without the painstaking, time-consuming ETL processes.
However, one size never fits all. Our co-development efforts aim to bring the Neo4j graph intelligence capabilities that users need to their Snowflake data estate, with deployment options suited for different use cases. The integrated portfolio includes Virtual Graph, ** Graph Analytics for** Snowflake, and the ** Aura Data Importer** for Snowflake.
Build knowledge graphs without moving Snowflake data with Virtual Graph
Build and query knowledge graphs directly over Snowflake data with zero data movement – translate Cypher to SQL automatically, preserve existing governance, and accelerate AI reasoning.
Virtual Graph for Snowflake, currently in Public Preview with planned GA on November 16, allows users to build a knowledge graph over an organization’s Snowflake data, enabling them to run Cypher queries against data in the data warehouse with zero data movement. Virtual Graph translates Cypher queries into deterministic SQL statements that are pushed to a Snowflake data warehouse, returning the results (not the data) as a graph for exploration and use within Bloom and Query, use by Aura Agents, and more. Cypher is GQL-compliant, which assures organizations requiring robust standards for query languages.
This brings the power of Neo4j’s AI-powered graph tools to an organization’s relational data core, leveraging existing data governance rules and the compute platform in which they’ve already invested. Virtual Graph provides AI-powered graph schema design against existing Snowflake data. Users can generate their first graph model instantly and adopt a working knowledge graph in minutes (not weeks), accelerating time to business value. With Cypher in their toolkit, users can craft queries to ask connected data questions like “if Warehouse X in Houston shuts down due to a storm, which pending orders need to be rerouted, through which secondary hubs, to hit guaranteed delivery dates?” or “what clinical outcomes were achieved by patients with similar genetic mutations, age profiles, and co-morbidities when given Treatment A vs. Treatment B?”. Users unfamiliar with Cypher can make use of the AI-powered Copilot to translate natural language to graph queries.
Beyond tried and true knowledge graph use cases such as customer/product 360, supply chain management, fraud prevention, and medical research, the exponential growth of agents and agentic tools in the market is accelerating the adoption of horizontal graph use cases with the capacity to improve agentic outcomes. LLMs offer a vast breadth of knowledge and astounding capabilities to generate multimodal content, but their lack of company-specific context and understanding can limit their usefulness and, therefore, their adoption at scale.
Recent research<sup>1</sup> has shown that organizations using Neo4j to provide connected business context to enterprise data sets can “improve LLM reasoning accuracy, deliver more consistent and explainable results, and reduce LLM hallucinations by 44%”.<sup>2</sup> The ability to visualize connections (among applications, metric definitions, departments, and the other history and nuances that make a company unique) grounds LLMs for optimized, org-specific outcomes. Jesús Barrera, AI Field CTO at Neo4j, shared his insightful vision and a helpful breakdown of related terms in his blog, The knowledge layer for enterprise AI. Virtual Graph for Snowflake empowers businesses to build the essential knowledge layer for AI reasoning, operating directly within their existing Snowflake infrastructure with minimal time to value.
Uncover new patterns in Snowflake data with Graph Analytics for Snowflake
Run on-demand graph analytics with SQL—create in-memory graphs from Snowflake tables, apply algorithms, and write results back for downstream use, all fully contained in Snowflake.
Graph Analytics for Snowflake, available today through the Snowflake Marketplace, helps businesses uncover new patterns in their data without ETL, using only SQL and elastic compute, all within the governance of their Snowflake AI Data Cloud. Users can employ Neo4j’s 65+ graph algorithms to derive deeper insights from their data, where it resides. The serverless, fully-managed application spins up on demand to run graph algorithms in memory using SQL, writes results back to Snowflake tables, and spins down again.
With Graph Analytics for Snowflake, users are charged only for the resources they consume, with no monthly minimum or idle fees. Business users can leverage Snowflake Cortex AI paired with Neo4j graph context to ask questions in natural language and receive relationship-aware responses. Graph Analytics for Snowflake delivers the fastest time to value for users who wish to extract insights from their comprehensive Snowflake data estate without time-consuming, expensive ETL. Key use cases include supply chain optimization (“what is the most optimal route through a supply chain network?”), community fraud detection (“are there groups of bad actors sharing identifiers?”), entity resolution for marketing (“find the unique users across all our customer channels for more effective campaigns”), and customer segmentation/next best action (“what should I recommend to a customer next?”).
Upgrade to AuraDB for real time use cases
Deploy native graph databases with millisecond latency and ACID compliance to power real-time, mission-critical applications.
Certain workloads – those requiring millisecond latency, CRUD operations, ACID transactions, or queries across data sources – remain best served by a native Neo4j graph database, available through either AuraDB (Neo4j’s fully managed graph database service) or our self-managed offering. To support real-time decision-making for use cases such as cyber threat analysis, next-best-action recommendations, network and IT root cause analysis, and live fraud detection, users can migrate hotspots (“subgraphs”) into a Neo4j database. The Import Service in Neo4j Aura supports Snowflake as a data source, facilitating migration of transactional, high-concurrency datasets into a native Neo4j graph database.
Graph intelligence in action
The expansion of graph capabilities within the Snowflake Data Cloud is a milestone for our partnership. Organizations are already using Neo4j Graph Analytics for Snowflake to turn novel data insights into business outcomes. For instance, Audience Acuity accelerated identity resolution across 2B+ interconnected data signals within their Snowflake data estate, reducing processing time from multiple weeks to under 24 hours.<sup>3</sup> Now, Virtual Graph for Snowflake enables in-place querying of Snowflake data as a graph in place, to over 12,000 organizations,<sup>4</sup> opening new possibilities for enterprise AI reasoning.
Explore how Neo4j technology complements your Snowflake Data Estate
Virtual Graph for Snowflake is now available in the Aura Console. Connect a Snowflake data warehouse, generate a model, and run your first Cypher query in minutes. Ben Lackey’s recent developer blog provides step-by-step guidance.
Graph Analytics for Snowflake is available today in the Snowflake Marketplace. Prash Medirattaa’s developer blog demonstrates the process.
AuraDB**,** Neo4j’s fully managed graph database, runs on AWS, Azure, and Google Cloud. Get started for free today!
Learn more
- **Virtual Graph for Snowflake** Public Preview[documentation](https://neo4j.com/docs/virtual-graph/self-managed/jdbc-drivers/snowflake/)
- Neo4j Graph Academy lab: [Neo4j Virtual Graph with Snowflake](https://graphacademy.neo4j.com/courses/virtual-graph-snowflake)
- Zero-Copy Graph Reasoning on Snowflake: Getting Started With Neo4j Virtual Graph blog
- Graph Analytics for Snowflakeuser manual
- Identify Bottlenecks and Model Disruptions using Neo4j Graph Analytics developer blog
- **AuraDB**[documentation](https://neo4j.com/docs/aura/)
- [Getting Started with Neo4j Aura](https://neo4j.com/blog/auradb/get-started-with-neo4j-auradb/) blog
1. IDC White Paper, *The Business Value of the Neo4j Graph Intelligence Platform* , May 2026.[↩︎](#7baa0adb-ee48-465d-876c-a665dc918c8f-link)
- New research finds enterprises earn 230% ROI with Neo4j Graph Intelligence Platform, May 2026.↩︎
- Audience Acuity Accelerates Identity Resolution to Under 24 Hours with Neo4j Graph Analytics for Snowflake
[↩︎](#551736d1-d414-4d48-af90-8371d7663819-link)
4. [About Snowflake](https://www.snowflake.com/en/company/overview/about-snowflake/) , 2026[↩︎](#757be617-c6f5-44a4-a7b8-b9c6589c5a42-link)