You incur Atlas Stream Processing costs for the following items:
Running Stream Processors
Data Transfer
VPC Peering
Private Link connections
Stream Processors
Atlas charges per hour per stream processor, where the hourly rate is determined by the tier of the stream processor. Atlas Stream Processing bills users for stream processors only when they are running. Billing is recorded in one-second increments.
Important
SP10 and SP30 processors operate and bill users
according to the legacy worker model. These processors
update to the per-processor pricing model on December 3rd,
2025. To learn more, see the worker model section of the
Atlas Stream Processing architecture overview.
The allocated vCPU and RAM vary depending on your stream processor tier:
Stream Processor Tier | vCPU | RAM | Bandwidth | Max Parallelism | Kafka Partition Limit |
|---|---|---|---|---|---|
| 0.25 | 512 MB | 50 Mbps | 1 | 32 |
| 0.5 | 1 GB | 125 Mbps | 2 | 64 |
| 1 | 2 GB | 200 Mbps | 8 | Unlimited |
| 2 | 8 GB | 750 Mbps | 16 | Unlimited |
| 8 | 32 GB | 2500 Mbps | 64 | Unlimited |
Note
Atlas doesn't charge for storage.
Data Transfer
Important
Updated Data Transfer Pricing
New data transfer pricing for Atlas Stream Processing takes effect on December 3, 2025 for S3-or-MongoDB-based sinks. Current data transfer pricing applies to usage prior to this date. The updated pricing varies by provider and transfer type, making intra-region and inter-region data transfer more affordable than the current pricing structure.
Atlas Stream Processing charges for all egress data transfer based on the transfer type and destination. Data transfer pricing varies by cloud provider and follows a tiered structure from lowest to highest cost. Data transfer charges per GB.
VPC Peering
Atlas Stream Processing charges for all VPC peering connections. To learn more, see Networking.
IMPORTANT: Atlas Stream Processing supports VPC peering connections only for AWS and Google Cloud workspaces.
Private Link
Atlas Stream Processing charges for all Private Link connections. To learn more, see Add a Kafka Private Link Connection.