Graph database biz Neo4j doubles total funding courtesy of $80m E-series splurge

Firm plans to stuff a sizeable chunk into AI, naturally

Graph database-flinger Neo4J has doubled its total funding after bagging $80m in E-round funding.

The biz, which is headquartered in Silicon Valley but has strong links to Europe via Swedish founder and CEO Emil Eifrem, launched in 2007.

To date, the most it has pulled in during a single funding round is £36m, having raised a total of $80m prior to today's funding round. This means today's E-series funding – led by UK-based One Peak Partners and Morgan Stanley Expansion Capital – doubles its total to date.

The firm didn't divulge its valuation, but the with a big name like Morgan Stanley in the picture, this puts it in good stead as it attempts to fend off new players in the graph database market, which has been predicted to grow from $821.8m in 2018 to $2.4bn in 2023.

Neo4j has regularly been ranked as the number one graph database firm on DB Engines, but new players, such as TigerGraph – which came out of stealth last September with a $31m series A round – are increasing the pressure.

There are also graph database products from more established vendors with larger footholds in the enterprise, for instance AWS's Neptune and Microsoft's Cosmos.

Eifrem acknowledged this, but said it had changed conversations during this funding round. Previously, investors queried the size of the market, but this time Eifrem claimed the focus had shifted to how Neo4j would maintain its position.

He told The Register that his answer was "a technology play", and that a "significant proportion" of the $80m funding would be funnelled into Neo4j's product to make it "faster, more scalable and easier to use" than competitor products.

He also argued that Neo4j's tech has "structural advantages" over some of the companies they are facing because it is platform neutral.

"You can run it on Amazon, on Azure, on GCP or in a data centre," he said. "That is something that one of the big platform players just won't do because their graph database business is part of a much bigger business."

The boss is also eyeing up artificial intelligence as a potentially lucrative business direction, and is pitching the idea that graph is crucial for AI to the enterprise, because it "gives colour" and context.

"Machine learning systems today do feature extraction – you extract these features from sample data and train model with that. What's missing, what you won't capture, is the connections," he said. Neo4j's argument is that ML systems trained on connected graph datasets bring in facts related to relationships, and so improve the predictions the models make.

But Neo4j can't yet publicly point to that many customer use cases – eBay's AI "shopbot" application, which aims to improve search on the online marketplace is one – and Eifrem acknowledges that there is still "a huge educational journey" to bring in more customers.

He also stressed that the firm was continuing its focus on its "bread and butter" business and transactional application, with its latest release, Neo4j 3.5, coming out last month.

New additions in this are full text search, which he said would be useful for natural language processing, and an official Neo4j driver for programming language Go. ®




Biting the hand that feeds IT © 1998–2018