Enabling AI requires creating connections between vast amounts of data. That's where technologies like graph databases come in.
Graph databases handle rapidly changing, interconnected data better than traditional databases designed to store strictly structured information. Of course, graph databases require management to take advantage of them. Many companies sell products for this purpose, and one of the leading vendors is Neo4j.
Neo4j's roots date back to the early 2000s, when founders Emil Eifrem, Johan Svensson, and Peter Neubauer discovered problems with traditional database technology. The trio developed a prototype of Neo4j, which later became the company's graph database management software of the same name.
“We came up with the idea for the first property graph database on a flight to Mumbai in 2000,” Eifrem told TechCrunch. “We sketched it on a napkin. I wish it was still there, but unfortunately it's gone.”
Neo4j was launched in Sweden in 2007, where Eifrem, Svensson and Neubauer were based at the time. In 2011, the company moved to Silicon Valley to raise venture funding.
Today, Neo4j's software allows businesses to build, tune, and deploy graph databases. Like other graph databases, Neo4j stores data as nodes, relationships, and properties. Nodes hold information about entities such as people or products. Relationships represent connections between nodes. Properties add further detail to nodes and relationships.
Diagram of Neo4j database and management tools. Image credit: Neo4j
Neo4j's graph database allows you to query data in a way that reflects how real-world entities are connected. This is a boon for AI. The data in a graph database is represented as a “knowledge graph” and is built around a context that allows AI to inform its output.
With the rise of AI, Neo4j has invested heavily in something called “GraphRAG,” a technology that allows AI to retrieve data from external sources. GraphRAG uses knowledge graphs to represent data and associated metadata in documents, potentially improving AI performance.
Neo4j also introduces a new vector search feature that captures relationships in your database based on items with similar characteristics. Vector search is useful for AI that needs to find similar text or files, make recommendations, or identify broad patterns.
The increased focus on AI support capabilities has benefited Neo4j. The company said its revenue was more than $200 million, double what it was three years ago, and that it expects to be cash flow positive “in the coming quarters.”
Neo4j accounts for 44% of the graph database market (according to a Cupole Consulting Group report), counts 84% of Fortune 100 companies including IBM and Walmart as customers, and plans to add more AI capabilities to its platform next year is.
“Companies are increasingly turning to AI to understand what it can do for their organizations. It needs to be explainable to a reasonable person,” Efrem said. “Our technology helps organizations achieve successful production deployments faster and more efficiently.”
Neo4j, valued at $2.2 billion, has 800 employees and 1,700 customers, plans to eventually go public. But for now, the focus is on growth. The company recently secured $50 million from Neotus Partners to “strengthen its balance sheet.” (To date, Neo4j has raised approximately $550 million in venture capital.)
Even if Neo4j waits years for an IPO, the graph database sector is likely to remain strong. According to Grand View Research, the graph technology market is expected to be worth $15.8 billion by 2030. Gartner also predicts that by 2025, 80% of data and analytics innovation will be done using graph technology.