Inside Infinigraph: how Neo4j scales a single graph

AI at massive scale needs a graph engine that does not blink at 100 TB. Enter Infinigraph.

I had a blast at GraphSummit, London, I spoke to Ivan Zoratti, VP Product Management, Neo4j to dig into the Infinigraph announcements and what they unlock for real workloads.

What we covered

* What Infinigraph is and who needs it now

* Property sharding in plain terms. How data spreads across nodes without losing graph semantics

* One engine for ops and analytics. How HTAP stays fast without starving either side

* Migration path for current Neo4j users. What carries over and what to plan for

* Where it shines at 100 TB and up. Boundaries, guardrails, and real results on speed and cost

* Timelines to ship and what this means for agentic AI next

Why it matters

Bigger graphs with lower latency change what agents can do. If you want real-time reasoning on live data, the storage and compute model must scale without falling apart.

Watch if you care about scale, cost, and a clean path from today’s Neo4j to what is coming next.

#Neo4j #GraphSummit #Infinigraph #GenAI #AgenticAI #GraphIntelligence #TheRavitShow