Welcome back to another episode of the EUVC Podcast. Today, Jeppe sits down with Emil Eifrem, founder & CEO of Neo4j, the world’s leading graph database and a core infrastructure layer for AI applications used by all 20 of the top US banks, 9 of 10 global pharma giants, and every major automotive OEM.
Emil recently announced a $100M global startup program to back founders building the next generation of AI-native products on top of graph technology, from knowledge graphs to hallucination-free LLMs.
We delve into why graph thinking matters now, how Neo4j came of age during the Panama Papers investigation, and why Europe is better positioned than people think to compete in the AI platform shift.
Here’s what’s covered:
02:00 The Panama Papers “Coming Out Party”
How journalists used Neo4j to uncover 7-layer-deep financial relationships invisible to traditional databases and why it triggered a wave of global adoption.06:40 Why Graphs Are the Missing Link for AI
Knowledge, meaning, context, and relationships: why LLMs without structured knowledge graphs hallucinate.08:50 The $100M Startup Program
Why Neo4j is returning to its roots to support AI-native founders and why the packaging for startups had to change.12:00 What Founders Get
Free Aura credits, dedicated graph engineers, joint GTM, and access to the world’s largest graph developer community.14:30 Early Traction: 300+ Startups in Weeks
Why early demand is far ahead of expectations and the kinds of companies applying.16:10 Community as a Strategic Moat
500+ annual global events, deep developer love, and why skill availability is now a CIO-level buying criterion.19:00 Building Deep Tech in Europe
Why Neo4j kept engineering in Europe, how the ecosystem matured, and what today’s founders can learn.22:00 Regulation & Competitiveness
Will Europe overregulate itself out of the AI race? Emil’s perspective on models vs infrastructure vs applications.23:40 The Future of AI Infrastructure
Why every company must rethink its stack and why the biggest threat is assuming your business will survive without change.
✍️ Show Notes
Neo4j: Europe’s Quiet AI Infrastructure Champion
Neo4j started as a fringe idea:
“What if we represent data as networks, not tables?”
A decade later, it’s the backbone of knowledge-driven AI, powering everything from fraud detection and drug discovery to supply-chain intelligence and autonomous systems. Neo4j now serves:
All 20 of the largest US banks
9/10 of the top global pharma companies
Every major global car manufacturer
Graph technology has become mission-critical as enterprises rush to build AI that understands, reasons, and retrieves real knowledge, not just probabilistic text.
The Panama Papers Moment
Neo4j’s breakout came when investigative journalists faced 1.6TB of unstructured, messy data, passports, scanned PDFs, and shell company paperwork. Traditional databases failed.
Graphs didn’t.
They revealed multi-hop relationships impossible to find in tables:
Addresses → People → Officers → Subsidiaries → Bank Accounts offshore.
The result:
The biggest global news story of 2016
A Pulitzer Prize
And a global wake-up call for banks and regulators: “How did journalists understand our customers better than we could?”
Neo4j became the default graph database for serious investigations, financial crime, and entity resolution.
Why Graphs Matter Now — Even More Than in 2016
LLMs are powerful but brittle.
They hallucinate. They forget. They confuse entities.
Graphs solve these problems by giving AI models structured memory, context, and ground truth.
Examples:
RAG systems with knowledge graphs drastically cut hallucinations
Context-aware agents can reason using real relationships
Enterprise AI systems can track provenance and explain decisions
Graphs are to AI what GPUs were to deep learning.
The $100M Startup Program: Betting on the Next AI Wave
Emil’s insight:
Startups will adopt new AI architectures much faster than banks and corporates.
But Neo4j had a problem. They had become an enterprise company:
Suits, not hoodies
$250k entry contracts
Long sales cycles
Heavy packaging
Great for Fortune 500s.
Terrible for founders.
So Neo4j is changing the model:
What Startups Get
1. Free Neo4j Aura credits
Run production graph databases in the cloud at zero cost.
2. Dedicated graph engineers
Architectural reviews, design patterns, and brainstorming features. A huge advantage when everything in AI is moving weekly.
3. Co-marketing & GTM
Exposure to the world’s largest graph community, speaking slots, blogs, and joint announcements.
The response?
300+ startups in under two months putting Neo4j on track to run one of the world’s largest AI startup cohorts.
Community as a Moat
Before AI became hot, Neo4j built a bottom-up community:
500+ in-person events per year (pre-pandemic)
Millions of downloads
Thousands of developers globally are trained in graph thinking
CIOs now consider this a feature:
“I can hire talent because you already trained them.”
This is something European deep-tech founders can learn from:
Communities compound.
And they are defensible.
Europe’s AI Advantage (and Weakness)
Emil’s take:
Europe can win, but not everywhere.
Where Europe can compete:
Infrastructure layers
Applied AI
Vertical AI apps
Tools around LLMs and agents
Where Europe may struggle:
Foundation models
Capital intensity is brutal. Mistral may be the exception, but it’s not yet the global leader.
Biggest risk?
Overregulation is killing momentum before startups scale.
The Bigger AI Threat
This shift is so fundamental that Emil no longer views Neo4j’s startup program as optional:
“If we don’t win this next platform shift, we may not have a business to protect.”
AI is not a feature.
It’s a replatforming of the entire software stack.
And graph-powered AI is the architecture shift happening beneath that.
💡 One-Liner Takeaway
The future of AI belongs to systems that understand context. And context lives in graphs.
