Introducing Aerospike Graph Database 3.0: Faster, simpler, and built for the terabyte scale era

Today, we’re announcing the release of Aerospike Graph Database 3.0, bringing major improvements across the three dimensions that matter most: developer ease with multi-property and native datetime support, 10x faster ingest performance, and up to 50% better cost efficiency through reduced storage footprint. Built for teams managing the most demanding graph workloads in identity resolution, fraud prevention, and real-time personalization, Aerospike Graph Database 3.0 is our biggest step forward in making high-scale, graph data fast, accessible, and affordable.

Easier to develop

Aerospike Graph Database 3.0 introduces two key capabilities that make modeling and querying graph data more natural and powerful:

Multi-property support

Vertices can now store multiple values under the same property key. This simplifies how you can model complex signal relationships, such as multiple timestamps, scores, or segments, without complex workarounds or nested structures.

Example: Store multiple tags on a user vertex

g.addV('User')
.property('T.id', 'user-123')
.property(VertexProperty.Cardinality.list, 'tag', 'frequent-buyer')
.property(VertexProperty.Cardinality.list, 'tag', 'mobile-user')
.property(VertexProperty.Cardinality.list, 'tag', 'high-LTV')

Query all tags for the user

g.V().has('id', 'user-123').values('tag')
// Output: ['frequent-buyer', 'mobile-user', 'high-LTV']

Or query Users with a tag

g.V().has(“tag”, “high-LTV”)
// Output: [user-123]

This is especially useful in AdTech, identity graphs, fraud, and personalization graphs, where entities accumulate multiple attributes over time.

Native Datetime support

Timestamps are now supported as a first-class property type. You can store, filter, and compare temporal data directly with no need to convert or encode datetime fields manually.

Example: Add a timestamp to a session vertex

g.addV('Session')
 .property('id', 'sessionid-123456')
 .property('startedAt', datetime('2025-07-08T14:22:00Z'))

Query sessions that started in the last 24 hours

g.V().hasLabel('Session')
 .has('startedAt', gt(datetime('2025-07-07T14:22:00Z')))

These capabilities make the platform more expressive and developer-friendly, especially for time-based and signal-rich domains.

Faster to load

Aerospike Graph 3.0 ships with a re-architected bulk loader that dramatically reduces ingest time, especially at large scale.

In production-scale benchmark tests, we loaded a terabyte of data in under three hours, down from over 32 hours with version 2.6. That’s a >10x performance gain, with the same infrastructure. 

This performance unlocks faster iteration and experimentation cycles, allowing teams to load large datasets quickly, validate changes, and refresh data frequently, without needing to scale up infrastructure or wait for a long time to see the results of their changes.

Cheaper to run

Aerospike Graph Database 3.0 also delivers substantial gains in storage efficiency and overall total cost of ownership.

Improvements to on-disk data layout reduce storage footprint by up to 50%, which means fewer nodes are required to store the same data. This directly reduces cluster size and infrastructure costs.

• Operate a terabyte-scale graph for a fraction of the cost of any other graph database

• Enjoy predictable cost scaling, with $/GB dropping as data volume increases

Combined with Aerospike’s Hybrid Memory Architecture and the ability to scale compute and storage independently via the Aerospike Graph Service, Aerospike Graph offers the most cost-efficient ways to run high-performance graph workloads, without the constraints of RAM-bound architectures.

Get started

Aerospike Graph Database 3.0 is:

• Easier to build on - with multi-property vertex support and native datetime types

• Faster to load - with >10x ingest speed for billion-scale datasets

• Cheaper to run -  with smaller clusters, lower storage footprint, and reduced total cost of ownership at scale 

If you’re solving large-scale graph problems that power real-time decisions, Aerospike Graph Database 3.0 is built for you.

→ [Read the release notes]