Amazon SageMaker
An integrated AWS environment for analytics, machine learning, generative AI, and governed access to enterprise data.
About Amazon SageMaker
Amazon SageMaker brings AWS machine learning, analytics, data processing, SQL, and generative AI development into a connected environment. Teams can build, train, customize, and deploy machine learning and foundation models with managed infrastructure, work in serverless notebooks, query data through a built-in SQL editor, and create data pipelines with familiar AWS services. SageMaker Unified Studio provides shared access to data and tools, while SageMaker Catalog supports discovery, governance, access control, and collaboration. Its lakehouse architecture connects Amazon S3 data lakes, Amazon Redshift data warehouses, and federated or third-party data sources. Amazon Q Developer adds natural-language help for finding data, writing code and SQL, troubleshooting, and guiding model development. Costs are usage-based across the SageMaker capabilities, AWS compute and storage, catalog requests, networking, and any third-party services used; selected setup and management requests are always free, and eligible AWS Free Tier benefits may apply.
Best for
Best use cases
Building and deploying machine learning or foundation models
Analyzing data across Amazon S3, Amazon Redshift, and federated sources
Creating governed data and AI projects in a shared development environment
Compare before choosing
Compare the total cost of all AWS services and resources used, not only the SageMaker interface
Assess whether AWS-native data, identity, and governance workflows match the existing stack
Check whether managed notebooks, SQL tools, model workflows, and catalog controls reduce tool switching
Features
Develop and deploy ML models
Prepare training data, develop machine learning or foundation models, run managed training and evaluation workflows, and deploy models for inference using SageMaker infrastructure. This supports teams that want one AWS workflow from experimentation through production operation.
Run analytics across data sources
Use managed notebooks and SQL tools to analyze data stored in Amazon S3, Amazon Redshift, and supported federated or third-party sources through a lakehouse architecture. Analysts and engineers can work with shared data without maintaining separate copies for every tool.
Govern collaborative data and AI work
Organize projects and assets, discover and share data and models, and apply fine-grained access policies through SageMaker Catalog. This gives data, analytics, and AI teams a common governance layer for collaborative work across the development lifecycle.
Pros
- Combines analytics, data processing, SQL, ML, and generative AI workflows within AWS
- Provides managed infrastructure for training, customizing, and deploying models at scale
- Applies shared discovery, access control, and governance to data, models, and AI assets
Cons
- Total cost depends on several separately billed AWS services and workload resources
- Quick setup can create additional networking charges that vary with account configuration
- Teams outside the AWS ecosystem may face integration and operational learning costs
Pricing Plans
Usage-based
- Charges follow the AWS services and resources used
- SageMaker Catalog usage is billed separately
- Compute, storage, networking, and third-party services may add charges
- Selected Unified Studio setup and management requests are always free
* Pricing information is for reference only. Please refer to the latest announcement on Amazon SageMaker's official website.