| Feature / Capability | Atlas Infinite | Atlas Core |
|---|---|---|
| Ideal workload profile | Most workloads, especially spiky, AI-driven, agentic, or high-growth applications where demand can change without warning. | Predictable, steady-state workloads with well-understood, consistent traffic patterns. |
| Compute and storage relationship | Compute and storage scale independently and rapidly, automatically absorbing sudden spikes in reads and writes. | Compute and storage are provisioned together in an attached configuration as fixed cluster sizes. |
| Storage capacity per jigh availability (HA) shard | Up to 128 TB per HA shard to support up to petabyte-scale clusters. | Up to 4TB per HA shard, sized for steady, well-scoped datasets. |
| Elasticity and scaling speed | Rapid vertical and horizontal scaling designed to absorb unexpected bursts in traffic. | Vertical and horizontal scaling supported, sized and executed as planned resize operations. |
| Security model | Multi-tenant architecture with dedicated compute in your VPC plus zero-knowledge storage where data and metadata are encrypted before reaching the shared layer. | Single-tenant clusters with strong isolation and standard Atlas encryption. |
| Background operations impact | All background operations, including backups, restores, replication, and maintenance, run on the storage layer, so they never compete with live traffic. | Background operations share cluster resources and should be scheduled around peak times. |
| Financial model | Available on Dedicated clusters. Storage is purely consumption-based, metered by actual data stored. Compute runs on a reserved instance tier, billed for the running shard, not query volume. | Provision-based: capacity reserved up front for consistent monthly charges. Free and flex tier clusters are also available. |
| Multi-cloud portability | Available on AWS today, with additional clouds planned. | Run a single cluster across AWS, Microsoft Azure, and Google Cloud. |
| Developer experience | Same MongoDB API, drivers, and tooling as Atlas Core, ready for AI and large-scale growth. | Familiar Atlas experience focused on steady-state workloads. |
| Limitations | Some Atlas features are not available during the Atlas Infinite public preview. See the full list in the documentation. | No limitations except for new Atlas features that may be in a public preview state. |
Atlas deployment options tailored for your needs
Give every workload the right foundation: Atlas Core for predictability, Atlas Infinite for AI-scale volatility.
Choose how to deploy your database
MongoDB Atlas offers a single database platform with two deployment options. Pick the deployment option that matches how your workloads behave today and where they need to go tomorrow.
Public preview
Atlas Infinite
Purpose-built for AI in production, autonomous agents, and applications facing growing and unpredictable demand. Absorb any spike in demand and scale from prototype to petabyte, across any cloud, without ever facing a disruptive rebuild.
- Independent scaling of compute and storage for rapid elasticity
- Seamlessly scale from prototype to petabyte
- Same Atlas UI, APIs, drivers, and operational tooling
Stable workloads
Atlas Core
Built for predictable workloads that need fixed capacity and budget clarity. Reserve exactly the resources your workload requires, delivering consistent, mission-critical performance you can plan around.
- Attached compute and storage for predictable capacity
- Multi-region clusters for global resilience
- A proven, mission-critical track record
Atlas Infinite for AI-scale growth
Demand is accelerating: bigger, more variable, harder to predict, and AI agents are pushing it further still. You need a database built for that reality, not stretched to accommodate it. That is Atlas Infinite.
EXTREME ELASTICITY
Infinite scale, zero waste
Atlas Infinite is purpose-built to support today's speed and scale of demand: more applications, shipped faster, at massive scale, always accelerating. It delivers the performance, responsiveness, and retrieval accuracy your AI agents demand.
- Peak performance and elasticity for growing workloads
- Retrieval accuracy and responsiveness for AI agents
- The same Atlas UI, APIs, drivers, and tooling
AI-scale, zero waste
Thousands of AI agents can generate unpredictable spikes in demand. Atlas Infinite automatically scales capacity up and down in real time, so you only use what you need.
Peak performance, no compromises
Atlas Infinite query is designed to maintain performance as your workloads grow, helping you scale without the infrastructure constraints that slow innovation.
Never outgrow your database
No matter how big your business gets, Atlas Infinite will grow with you. Scale on your terms, when you need more performance, not when you run out of room.
Zero-knowledge security
Cloud-scale efficiency without compromising your security posture. Atlas Infinite encrypts your application's data before it hits storage, keeping the storage layer blind.
Same Atlas experience
No new system to learn, no disruptive migration, no code changes. Just a new way to deploy Atlas. It runs on the same drivers, APIs, security, and tooling you already use today.
Atlas Core for steady-state apps
Choose Atlas Core for predictable workloads that need fixed capacity, budget clarity, and consistent, mission-critical performance.
PERFORMANCE AND CONTROL
Fixed resources, full control
Atlas Core provides steady, business-critical workloads a dedicated foundation of fixed resources, predictable performance, and clear capacity planning. The managed Atlas experience keeps upgrades, backups, and monitoring simple while teams control cluster sizing and regional placement.
- Dedicated resources for predictable workloads
- Clear capacity and budget planning
- Multi-region deployment and recovery
Reserve the resources your workload requires
Dedicated compute and attached NVMe storage give steady-state applications consistent capacity, budget clarity, and a predictable operational baseline.
Design resilience across the regions you serve
Deploy multi-region clusters to place data near users, support failover, and meet latency or residency requirements.
Performance proven at enterprise scale
Trusted by 70,000+ customers and 75%+ of the Fortune 100, Atlas Core is the mission-critical system of record that enterprises build on.
Which deployment option fits your workload?
Most workloads benefit from the elasticity and scale of Atlas Infinite, especially unpredictable, high-growth, AI, and agentic workloads. Atlas Core remains the right fit for steady-state workloads that need fixed capacity and tight cost planning.
$1.1MTotal cost savings
DATABASE MODERNIZATION
“When we moved to MongoDB Atlas, we saw incredible cost savings and performance lifts. These wins empowered us to continue to roll this solution out, which led to more savings, which led to more performance improvements.”
Robert Wise
Chief Information Officer, TriZetto
Chief Information Officer, TriZetto
$1.1MTotal cost savings
DATABASE MODERNIZATION
“When we moved to MongoDB Atlas, we saw incredible cost savings and performance lifts. These wins empowered us to continue to roll this solution out, which led to more savings, which led to more performance improvements.”
Robert Wise
Chief Information Officer, TriZetto
Chief Information Officer, TriZetto
$600MARR achieved
ENTERPRISE AI & SCALABILITY
"For any infrastructure the best outcome is that the engineering team doesn't have to worry about it because it just runs. You focus on the innovation with the models because you don't have to worry about the data layer—and the data layer is MongoDB."
Alex Holt
Field CTO, ElevenLabs
Field CTO, ElevenLabs
3xImprovement in search quality score
GEN AI & AI TALENT MATCHING
“We are still sticking with MongoDB because it is evolving with the technologies. It’s providing all the solutions we need.”
Pratham Rasal
Senior Software Engineer, Mercor
Senior Software Engineer, Mercor
70Products on data graph
Unified Data Layer
"Having a partnership with MongoDB from day one allowed us to build the employee graph and to build a capability set so that each successive product we were creating was a lot easier."
Ankur Bhatt
Former Head of AI, Rippling
Former Head of AI, Rippling
FAQ
Atlas Infinite is a new deployment option within MongoDB Atlas, not a new database to adopt. It runs the same MongoDB database on the same document model, with the same APIs, drivers, security protocols, and operational tooling you use today. No code changes, no application rewrites, no migration project.
What changes is how the database runs underneath. Atlas Core uses a traditional attached-storage architecture with fixed, dedicated resources. Atlas Infinite decouples compute and storage so each scales independently, which is what delivers rapid elasticity, up to 128TB per HA shard, and zero-knowledge security. Both are deployment options within the Atlas platform, and you choose the right one per workload.
Use Atlas Infinite when you expect volatile demand, rapidly growing datasets, or AI- and agent-heavy workloads that need elastic compute and high storage ceilings. Use Atlas Core when workloads are steady, you prefer pinned resources and predictable spend, and you want dedicated hardware for transactional and line-of-business services.
As of MongoDB’s initial Public Preview launch, Atlas Infinite is available on AWS, with support for additional cloud providers coming soon.
No. Atlas Infinite uses the same MongoDB API, drivers, and operational tooling as Atlas Core, so applications that already connect to MongoDB do not require application code changes. The main differences are architectural—how compute and storage scale and how capacity is managed—not how your application talks to the database.
Atlas Infinite splits pricing by resource: storage is purely consumption-based, metered by the actual data you store, so you're never paying for idle capacity. Compute, however, is billed at a reserved instance tier for your running cluster, not by query volume or usage. Atlas Core follows a provision-based model for both compute and storage, where you reserve fixed cluster sizes for a predictable monthly bill. Both deployment options use the same Atlas billing experience; the difference is how tightly spend tracks real demand.
Atlas Infinite runs compute in a per-customer VPC and encrypts data and metadata at the compute layer before anything reaches the shared storage system. That means other tenants cannot access your environment, and the underlying cloud platform cannot see your collections, contents, or access patterns. You get the efficiency of shared storage while aligning with strict zero-trust and data sovereignty requirements.
Start your next workload with the right Atlas deployment
Choose Atlas Infinite for volatile or AI-heavy workloads, and Atlas Core for steady services. Give every new application a foundation that’s built for it.
WHAT YOU GET FROM DAY ONE:
- 125+ regions worldwide
- Vector search and full-text search
- Intelligent query generation
- Full-lifecycle encryption
- Languages, frameworks, and tools of your choice