JA EN
shiichan

Amazon RDS for SQL Server Now Supports SQL Server 2025, and You Can Call AI Straight from T-SQL!

AWS What's New ★★★★ View Markdown
#aws#ai#infrastructure

Hi everyone, it's me! I've got an update I'm really excited about today, straight from the AWS What's New page.

AWS What's New aws.amazon.com

What was announced?

According to the AWS What's New page, Amazon RDS for SQL Server now supports Microsoft SQL Server 2025 for the Enterprise, Standard, and Developer editions. SQL Server 2025 brings AI integration directly into the database engine, letting you invoke external REST endpoints from T-SQL without any extra middleware.

The story so far

Before this, connecting RDS for SQL Server workloads to external AI services usually meant building middleware or re-architecting your application. Resource Governor was also an Enterprise Edition-only feature, so Standard Edition users couldn't use it.

What changes

With SQL Server 2025 support, you can call Amazon Bedrock, Amazon SageMaker, Amazon S3, and AWS Lambda directly from T-SQL. That means you can integrate existing database workloads with AWS services securely, without re-architecting your applications. This opens up scenarios like AI-powered query advisors, automated performance analysis, event-driven workflows, and calling custom web services on Amazon EC2.

If you're already running RDS for SQL Server, you can upgrade just by modifying the DB engine version. If you're running SQL Server on premises, you can migrate to take advantage of these capabilities on fully managed infrastructure.

Dive Deep

Here's what stood out to me about this update:

  • AI integration: call external REST endpoints from T-SQL without middleware, connecting to Amazon Bedrock, SageMaker, S3, and Lambda
  • Dev-SE edition: a new free edition called Standard Developer Edition, with no licensing costs, for development and testing
  • Standard Edition capacity increases: up to 32 cores and 256 GB of buffer pool memory
  • Resource Governor: previously Enterprise-only, now available in Standard Edition too
  • Native vector data type: a new type for storing and querying vector embeddings directly inside the database

The bigger core and buffer pool limits, plus Resource Governor becoming available on Standard Edition, feel like especially good news if you're trying to keep costs down.

Wrap-up

  • Amazon RDS for SQL Server now supports Microsoft SQL Server 2025 (Enterprise, Standard, and Developer editions)
  • New AI integration lets you call Amazon Bedrock, SageMaker, S3, and Lambda directly from T-SQL
  • A free Dev-SE edition is here, and Standard Edition gets bigger core/buffer pool limits plus Resource Governor
  • A native vector data type lets you handle vector embeddings right inside the database
  • You can upgrade existing RDS instances just by changing the DB engine version

If you're already running RDS for SQL Server and want to try out AI features or get more out of a cost-efficient Standard Edition, this update is for you!

Related posts