MongoDB Atlas Dedicated cluster pricing starts at $0.08 per hour for an M10 on AWS US East and climbs to $33.30 per hour for an M700 on AWS Frankfurt, a 415x range that puts cluster tier selection above every other cost lever. Realised enterprise Atlas discounts at $500,000 annual commit land at 10 to 18 percent. At $2 million commit the band moves to 20 to 30 percent. At $5 million it reaches 28 to 40 percent. This guide covers Atlas Dedicated cluster pricing, Search, Vector Search, Stream Processing, and Atlas Enterprise Advanced, plus the six negotiation levers that move a MongoDB contract beyond the public sticker.
Atlas Dedicated cluster pricing
MongoDB Atlas charges per cluster hour. Cluster price is a function of tier (M10 through M700), region, cloud, and storage class. The cluster includes three replica-set nodes by default. Production deployments typically run M30 or higher.
| Cluster tier | vCPUs / RAM (per node) | AWS US East per hour | AWS Frankfurt per hour | Approx monthly (730 hrs) |
|---|---|---|---|---|
| M10 | 2 / 2 GB | $0.08 | $0.09 | $66 (US East) |
| M20 | 2 / 4 GB | $0.20 | $0.22 | $146 |
| M30 | 2 / 8 GB | $0.54 | $0.60 | $394 |
| M40 | 4 / 16 GB | $1.04 | $1.16 | $759 |
| M50 | 8 / 32 GB | $2.06 | $2.30 | $1,504 |
| M60 | 16 / 64 GB | $4.12 | $4.60 | $3,008 |
| M80 | 32 / 128 GB | $8.24 | $9.20 | $6,015 |
| M140 | 48 / 192 GB | $12.36 | $13.80 | $9,023 |
| M200 | 64 / 256 GB | $16.48 | $18.40 | $12,030 |
| M300 | 96 / 384 GB | $24.72 | $27.60 | $18,046 |
| M700 | 128 / 768 GB (NVMe) | $29.97 | $33.30 | $21,878 |
Cluster pricing includes the underlying compute, the standard EBS gp3 or Azure Premium SSD storage at a baseline IOPS rate, replica-set high availability, automated backups (point-in-time recovery to 24 hours, 7-day rolling snapshots), and standard support. The cluster price does not include cross-region replication, additional analytic nodes, BI Connector, premium support, or any of the Atlas add-on services.
Atlas Search and Vector Search
Atlas Search and Atlas Vector Search run on dedicated Search Nodes, sized independently of the database cluster. Search Nodes are priced per node hour, with vector search workloads commonly requiring higher-memory nodes than text search.
| Search Node tier | vCPUs / RAM | AWS US East per hour | Typical use case |
|---|---|---|---|
| S10 | 2 / 8 GB | $0.10 | Lightweight text search |
| S20 | 2 / 16 GB | $0.20 | Text search with sort |
| S30 | 4 / 32 GB | $0.40 | Vector search up to 5M vectors |
| S40 | 8 / 64 GB | $0.80 | Vector search 5 to 25M vectors |
| S50 | 16 / 128 GB | $1.60 | Vector search 25 to 100M vectors |
| S60 | 32 / 256 GB | $3.20 | Production-scale vector search |
Vector Search has become a fast-growing MongoDB line item as enterprise RAG deployments scale. The cost trap is dimensional drift: a 1,536-dimension OpenAI embedding workload sizes search nodes 3 to 5x larger than a 384-dimension sentence-transformer workload at the same vector count. Embedding model choice has a direct line to MongoDB monthly bill.
Stream Processing, Charts, App Services
Atlas Stream Processing is billed per Stream Processing Instance hour ($0.08 to $0.96 depending on tier) plus per-document processed ($0.000004 to $0.0001). Atlas Charts is included up to 100 MB monthly data scanned, then $0.50 per GB scanned. Atlas App Services (Functions, Triggers, GraphQL, Device Sync) bills per request, per compute-second, and per sync data transferred. App Services typically adds 5 to 15 percent on top of the database cluster cost for application backend use cases.
The line item that surprises customers: Cross-region data transfer between Atlas regions runs $0.02 per GB on AWS and $0.04 per GB on Azure. A multi-region active-active cluster moving 5 TB per month between regions adds $1,200 to $2,400 per month that did not appear on the cluster pricing page.
Atlas Enterprise Advanced
MongoDB Atlas Enterprise Advanced is the bundled tier that adds Advanced Security (LDAP integration, customer-managed encryption keys, audit logging, field-level encryption), Advanced Operations (private endpoints, network peering), and Enterprise Support. It runs as a premium on top of the cluster price, typically 30 to 45 percent on the underlying cluster cost.
For regulated workloads (financial services, healthcare, EU public sector), Enterprise Advanced is not optional. Customer-managed encryption keys, audit logging, and Private Endpoint connectivity are routinely required by infosec review. For non-regulated workloads, the Enterprise Advanced premium is negotiable down to 15 to 25 percent on the underlying cluster cost at $2 million commit and above.
Discount bands and commit structures
MongoDB Atlas does not publish enterprise discount bands. Realised discounts observed across advisor-led MongoDB negotiations during 2024 to 2026:
| Annual commit (TCV) | Cluster discount | Search Node discount | Enterprise Advanced premium |
|---|---|---|---|
| $100K to $500K | 5 to 12 percent | 0 to 5 percent | 35 to 45 percent |
| $500K to $2M | 12 to 20 percent | 5 to 12 percent | 25 to 35 percent |
| $2M to $5M | 20 to 30 percent | 12 to 20 percent | 15 to 25 percent |
| $5M+ | 28 to 40 percent | 20 to 30 percent | 10 to 20 percent |
MongoDB Atlas commit contracts run 1, 2, or 3 years. Two-year terms typically add 3 to 5 percentage points over one-year. Three-year terms add another 2 to 4 points. Unused commit carries forward inside the contract term but expires at term end. The carry-forward cap is negotiable; the default is 25 percent of the contract year.
Six negotiation levers
Beyond the headline cluster discount, six contract levers move the MongoDB Atlas deal materially:
One: cross-region data transfer included in commit at a defined per-GB rate, rather than billed at on-demand. For multi-region deployments this is a five to six figure annual saving.
Two: Search Node discount aligned to cluster discount. MongoDB's default position is to discount cluster aggressively and protect Search Node pricing. Push for parity.
Three: Enterprise Advanced premium capped as a percentage of cluster spend rather than charged as a flat add-on. This caps the premium even as cluster cost grows.
Four: Free-of-charge non-production clusters (dev, staging, sandbox) up to a defined cluster-hour cap. MongoDB will normally agree this at $1 million commit and above.
Five: True-down right of up to 20 percent at annual anniversary, capped at the commit reduction floor. Without this, all commit risk sits with the customer.
Six: Price-protection clause locking cluster, Search Node, and Stream Processing rates for the contract term. MongoDB raised list prices materially in 2023 and 2024; price protection is now a standard ask.
Atlas vs Enterprise Advanced self-managed
MongoDB Enterprise Advanced (the self-managed product, sometimes confused with Atlas Enterprise Advanced) is the alternative for customers who want to host MongoDB themselves. Pricing is per node per year, with named SKUs by cluster size. For very large deployments where the underlying cloud or hardware cost is already heavily discounted, self-managed Enterprise Advanced can run 20 to 35 percent cheaper than Atlas at the same node count. Most enterprise deployments still favour Atlas because of the operational savings, but the comparison should be modelled before signing a multi-year Atlas commit.
For the broader data platform context, see our Snowflake vs Databricks vs BigQuery comparison, Databricks pricing pillar, and Snowflake pricing pillar. For the underlying cloud commit interaction, see our AWS EDP negotiation and Azure MACC versus CTP. For broader cost controls, see our cloud cost optimization and cloud contracts guide. For engagement on MongoDB negotiation, see our cloud contract negotiation service.
Backup, point-in-time recovery, and snapshots
Atlas includes continuous backup with point-in-time recovery to 24 hours plus 7-day rolling snapshots in the base cluster price. Extended retention is a separately billed feature. Continuous backup retention beyond 24 hours runs $0.50 per GB per month for the additional retention. Cluster snapshot retention beyond 7 days runs $0.10 per GB per month. For a 1-TB cluster running 30-day point-in-time recovery, the additional backup cost approaches $300 per month, which finance teams routinely miss in the cluster cost projection.
Cross-region snapshot replication adds another line item: $0.02 per GB on AWS replication and $0.04 on Azure replication, plus storage cost in the destination region. For disaster recovery deployments with 4-hour recovery point objective targets, the cross-region replication cost commonly equals 15 to 25 percent of the primary cluster cost.
Federated and multi-region deployments
MongoDB Atlas Global Clusters provide multi-region active-active deployment with location-aware shard placement. Pricing is the sum of the cluster cost in each region plus inter-region data transfer plus federated query traffic. For a three-region active-active deployment (US East, EU West, AP Southeast), the cluster cost is approximately three times the single-region cost, plus 8 to 18 percent for inter-region traffic depending on workload pattern.
Atlas Data Federation is a separate billed product that allows queries across Atlas clusters and S3 buckets without copying data. Federation runs at $5 per TB of data scanned, charged on top of the source cluster cost. For analytics workloads where data lives in both Atlas and S3, Federation can replace ETL pipelines, but the per-TB scan rate accumulates quickly on un-optimised queries.