Senior Director, Product Management, Neo4j
4 min read
AuraDB Business Critical and Virtual Dedicated Cloud now offer High Memory instances with up to 2TB of RAM on both AWS and Google Cloud. We’ve also raised the maximum storage to 5TB on every cloud.
High Memory instances give your most demanding workloads the memory to stay fast and the storage to keep growing. Teams can run graphs of this size in a fully managed service instead of splitting them apart or operating their own infrastructure.
Why memory matters for large graphs
Some graphs are simply enormous. Knowledge graphs grow with every document, embedding, and revision. Fraud and AML graphs hold years of transactions. Identity resolution and customer 360 graphs absorb every interaction. Digital twins ingest a constant stream of telemetry.
Memory matters more for these workloads than for most databases. Graph queries are rarely simple lookups. They traverse relationships, evaluate paths, and retrieve context across connected entities, and they’re fastest when the working set of the graph (the frequently accessed nodes, relationships, and indexes) stays in memory. When a large graph spills to disk, multi-hop traversals and deep path queries are the first to feel it.
Until now, running a graph this size meant hosting Neo4j Enterprise Edition on your own infrastructure. AuraDB now handles workloads up to 2TB of RAM, fully managed.
What’s new
Generally available on AWS. High Memory instances launched on Google Cloud earlier this year, where customers are already taking them into production. AWS joins today.
Four new instance sizes on each cloud. Previously, 512GB was the largest size you could create yourself. High Memory adds:
| AWS (RAM) | Google Cloud (RAM) | Storage range | | 768 GB | 744 GB | 1.5 to 5 TB | | 1,024 GB | 976 GB | 2 to 5 TB | | 1,536 GB | 1,433 GB | 3 to 5 TB | | 2,048 GB | 1,952 GB | 4 to 5 TB |
*Each size comes with a matching vCPU allocation. Ask your account team for the options in your region.*5TB storage on every cloud. We’ve raised the maximum storage for AuraDB Business Critical and Virtual Dedicated Cloud from 2TB to 5TB on AWS, Google Cloud, and Azure. The increase applies to instances from 192GB of RAM and up, within a 1:16 RAM-to-storage ratio.
Where High Memory instances fit
Large knowledge graphs and GraphRAG. Keep documents, embeddings, and the relationships between them on one managed instance, with the memory to retrieve context quickly as the graph grows.
Fraud and financial crime. Run deep path queries across years of transactions without watching latency climb as the data does.
Identity resolution and customer 360. Resolve and traverse every interaction across a highly connected graph, in real time.
Migrations from self-managed. Move your largest existing deployments to AuraDB without cutting the graph into pieces to make it fit.
Built for production
Same workflows. Create, resize, , clone, and snapshot from the Aura API or Console, like any other AuraDB instance.
Works with your existing features. High Memory instances include the operational, scale, and security features of your tier, as they do on any other AuraDB instance, from secondaries and Multiple databases to SSO and Private Link.
Memory and storage set separately. Set them at creation or on resize, so a storage-heavy graph doesn’t force a compute upgrade.
Familiar billing. RAM is billed per GB-hour and includes disk at a 1:2 RAM-to-storage ratio. Additional storage is billed per GB-hour at the standard rate. Secondaries and d instances are charged at the usual rates.
Automatable. Create a High Memory instance, or grow its storage, with the same API calls you already use.
Availability
High Memory instances are available to AuraDB Business Critical and Virtual Dedicated Cloud customers on prepaid terms. Because they depend on regional cloud capacity, we confirm availability before enabling them for your project. Availability varies by region, so tell us early where you need to run.Don’t see the size you need? The sizes above are the ones you can request today. If your workload needs a different mix of RAM and storage, another cloud, or a path off a Neo4j Enterprise Edition deployment you host yourself, we can help. We work with our cloud provider partners on capacity and infrastructure planning, and will match a configuration to your workload and region. Tell us what you’re running and we’ll take it from there.
Get started
- In the Aura Console , open the sizing menu when creating an instance and selectRequest larger instance . You can also contact your Neo4j account team.
- We’ll review your workload and agree on the configuration with you.
- Once your project is enabled, create and manage High Memory instances from the Aura API or Console.