Amazon SageMaker is AWS's service for data, analytics and AI in one place. SageMaker AI trains, customizes and deploys machine learning models and foundation models on managed infrastructure, with HyperPod, JumpStart and MLOps tools. SageMaker Unified Studio adds notebooks with a built-in AI agent, a SQL editor and data processing, while a lakehouse on Apache Iceberg unifies data in Amazon S3 and Amazon Redshift. SageMaker Catalog, built on Amazon DataZone, handles governance and access. Billing follows each AWS service you use, and a Unified Studio free tier covers core setup and management requests.
It is built for data scientists, machine learning engineers, data engineers, analysts and platform teams at organizations that already run on AWS. A data team can build models and pipelines in one environment, while governance staff define fine-grained permissions that apply across analytics and AI tools. Enterprises with strict security needs use the catalog for classification and access policies, and generative AI developers can build applications on their own data. Customers include Toyota, Charter Communications, Lennar, Carrier and NatWest Group. Teams without AWS experience face a steeper start because costs and setup span several services.
Typical workloads include training and deploying a prediction model at scale, building a generative AI application tailored to company data, running SQL analytics across lake and warehouse data, and preparing data with processing jobs. Near real-time zero-ETL integrations bring operational databases and applications into the lakehouse, and federated queries read third-party sources in place. An analyst can ask Amazon Q Developer to find data, generate SQL or build pipeline jobs in natural language. Teams share data, models and applications as governed assets through projects and domains, which suits organizations moving from experiments to production.
