Pricing overview
Amazon SageMaker supports geospatial machine learning (ML), making it easier for data scientists and ML engineers to build, train, and deploy models using geospatial data. You pay only for what you use.
Four components that determine your bill:
- a monthly user fee for visualizing and collaborating on
geospatial workloads, accessing the geospatial data catalog, and using
the purpose-built geospatial container - charges for compute hours consumed for Amazon SageMaker Studio, Processing jobs, or running fully-managed geospatial operations
- charges for storage based on the size of data imported or cached
- charges per request for geospatial operations
Monthly user fee
You pay a monthly user fee of $150 per user for access to the SageMaker Studio notebook pre-loaded with the geospatial image for purpose-built geospatial visualization and integrated collaboration features. This monthly use fee also enables you to access a geospatial data catalog, and a purpose-built geospatial container that runs on SageMaker Processing.
Compute charges
- You pay based on the type of compute used. The SageMaker geospatial
- capability supports a choice of interactive CPU
- (ml.geospatial.interactive ) and GPU compute instances, and SageMaker
- Processing instances.The SageMaker geospatial capability offers the following compute instances:
- Launch an interactive geospatial notebook (ml.geospatial.interactive) to configure and arrange your workflowRun processing jobs (ml.geospatial.jobs) to analyze satellite imagery or other geospatial dataRun off-the-shelf, purpose-built ML models (ml.geospatial.models) on satellite imageryLaunch Amazon SageMaker Processing jobs to run large-scale geospatial worklodsYou are charged different rates for each type of compute instances used.
Storage charges
You pay storage charges based on the total amount of storage used per month (GB) and the time for which you’ve used storage.
Geospatial operations
You pay per request for geospatial operations. You are separately billed for each type of geospatial operation request your application makes to the service. You can convert latitude and longitude coordinates to a readable street address or place name using the reverse-geocoding operation. You can also snap GPS traces recorded during a drive to the OpenStreetMap road and path network using map matching.
SageMaker Free Tier summary for geospatial ML
As part of the AWS Free Tier, you can get started with SageMaker geospatial capabilities for free. The free tier starts from the first month when you create your first SageMaker resource. The Free Tier lasts 60 days and includes no monthly user fee, 10 free ml.geospatial.interactive compute hours, and up to 10 GB of free storage. After the 60-day free trial period is complete, or if you exceed the Free Tier limits defined above, you pay for the components outlined in the pricing table below.
SageMaker Free Tier summary for Geospatial ML
As part of the AWS Free Tier, you can get started with SageMaker geospatial capabilities for free. The free tier starts from the first month when you create your first SageMaker resource. The Free Tier lasts 60 days and includes no monthly user fee, 10 free ml.geospatial.interactive compute hours, and up to 10 GB of free storage. After the 60-day free trial period is complete, or if you exceed the Free Tier limits defined above, you pay for the components outlined in the pricing table below.
Pricing table
SageMaker geospatial capabilities also supports G5 and Processing instances. For details on G5 and Processing instance types and pricing refer to the Sagemaker pricing page.
Pricing examples
Example 1: Using SageMaker geospatial capabilities notebook and geospatial operations
You have a team of one data scientist and one geospatial analyst who want to try the SageMaker geospatial capabilities. Let’s say one of them builds an ML model to predict crop yield, for which they access Landsat 8 data for farm fields in Utah. The geospatial analyst wants to plot the crop yield predictions on the satellite image, explore the results, and analyze them. Throughout the analysis (after consuming the Free Tier resources), the team is logged in to SageMaker geospatial capabilities notebook for 10 hours for the month. During the month, the data scientist runs geospatial operations as jobs for 20 hours and generates 30 GB of data, which is stored for a month. The bill at the end of the month would be calculated as follows:
Monthly user fee = $150 * 2 = $300
Compute charges for ml.geospatial.interactive: $1.20 * 10 = $12.00
Compute charges for ml.geospatial.jobs: $0.40 * 20 = $8.00
Storage charges: $0.023 * 30 = $0.69
Total = $300 + $12.00 + $8.00 + $0.69 = $320.69
Example 2: Using SageMaker geospatial capabilities notebook, geospatial operations, and models
Let’s say you have a team of two data scientists and eight decision makers who want to use a model with SageMaker geospatial capabilities. The data scientists access 5 years of Sentinel-2 data for London and New York, run the SageMaker geospatial capabilities–provided model on Sentinel-2 data, view the model predictions, and share them with the decision makers. During the process (after consuming the Free Tier resources), the team is logged in to the SageMaker geospatial capabilities notebook for a total of 100 hours for the month, runs 220 hours of jobs to preprocess data, runs a prebuilt landcover segmentation model for 20 hours, and generates 200 GB of data, which is stored for a month. The bill at the end of the month would be calculated as follows:
Monthly user fee = $150 * 10 = $1500
Compute charges for ml.geospatial.interactive: $1.20 * 100 = $120.00
Compute charges for ml.geospatial.jobs: $0.40 * 220 = $88.00
Compute charges for ml.geospatial.model: $2.50 * 20 = $50.00
Storage charges: $0.023 * 200 = $4.60
Total = $1500 + $120.00 + $88.00 + $50 + $5.20 = $1763.20
Additional pricing resources
Easily calculate your monthly costs with AWS
Contact AWS specialists to get a personalized quote