← Back to all announcements
★★★★☆ 07/07/2026

Amazon RDS for Oracle now supports Oracle Database 26ai

Run GenAI, vector search, and natural language SQL natively inside Oracle—no extra infrastructure, no data movement required.

View original announcement →

Visual Summary

graph TD A{{RDS for Oracle 26ai}}:::announced B(Amazon Bedrock):::compute C([Select AI - NL to SQL]):::feature D([AI Vector Search]):::feature E([RAG from SQL]):::feature F((Database Users)):::external G(VPC PrivateLink):::compute H([JSON Duality Views]):::feature I([SQL Property Graphs]):::feature F ==>|"natural language"| A A ==>|"invokes models"| B A -->|"generates SQL"| C A -->|"vector embeddings"| D D -->|"grounds responses"| E E -->|"augments"| B A -.->|"private connectivity"| G G -->|"port 443"| B A -->|"dual access"| H A -->|"graph analytics"| I classDef announced fill:#ff9900,stroke:#ec7211,color:#fff,font-weight:bold classDef compute fill:#e3f2fd,stroke:#1565c0,color:#1565c0 classDef storage fill:#e8f5e9,stroke:#2e7d32,color:#2e7d32 classDef feature fill:#fff3e0,stroke:#e65100,color:#e65100 classDef external fill:#f5f5f5,stroke:#616161,color:#616161

What's New

Amazon RDS for Oracle now supports Oracle Database 26ai, Oracle's latest Long Term Support (LTS) release, bringing native generative AI capabilities directly into the managed relational database service. The integration includes Amazon Bedrock connectivity for invoking foundation models such as Anthropic Claude, Amazon Nova, and Meta Llama directly from SQL, alongside Oracle's Select AI feature for natural language-to-SQL query generation. Customers can create new 26ai instances or upgrade from existing Oracle Database 19c or 21c container databases (CDBs).

How It Works

  • Amazon Bedrock Integration: RDS for Oracle 26ai connects to Amazon Bedrock runtime endpoints, allowing database users to invoke foundation models (text generation, summarization, embedding generation) directly via SQL using the DBMS_CLOUD and DBMS_CLOUD_AI PL/SQL packages, without extracting data from the database.
  • Select AI: Oracle's Select AI feature translates natural language prompts into SQL queries and executes them, enabling non-technical users and developers to query relational data conversationally using the connected Bedrock LLM.
  • AI Vector Search: Vector embeddings can be stored alongside relational data within the same Oracle database, enabling semantic similarity searches and hybrid (keyword + vector) searches without provisioning a separate vector database.
  • RAG from SQL: Retrieval-Augmented Generation pipelines can be implemented entirely in SQL using Oracle AI Vector Search, grounding LLM responses in enterprise data that stays within the database boundary.
  • JSON Relational Duality Views: The same underlying relational data can be accessed and mutated as either JSON documents or relational tables simultaneously, eliminating the need to maintain separate data models.
  • SQL Property Graphs: Graph relationships can be defined and queried in-database using SQL/PGQ syntax, enabling graph analytics without a dedicated graph database.
  • Upgrade Path: Existing Oracle 19c or 21c CDB instances can be upgraded to 26ai via the Modify DB Instance workflow by selecting engine version 26.0.0.0; non-CDB 19c instances must first be converted to CDB architecture before upgrading.
  • Network Connectivity: Bedrock integration requires the RDS instance to reach bedrock-runtime.<region>.amazonaws.com on port 443, achievable via a VPC interface endpoint (AWS PrivateLink, recommended) or internet gateway.

Why It's Important

  • AI without data movement: Generative AI and RAG capabilities are executed in-place within the Oracle database, eliminating ETL pipelines, reducing latency, and keeping sensitive data within the governed database boundary.
  • Productivity gains for developers and analysts: Select AI removes the SQL expertise barrier, allowing business users to query enterprise data using plain English and accelerating developer workflows.
  • Consolidation of data infrastructure: AI Vector Search removes the need for a separate vector database (e.g., a standalone pgvector or Pinecone instance), reducing architectural complexity and operational overhead.
  • LTS release stability: Oracle Database 26ai is a Long Term Support release, meaning customers receive extended patch support and can plan multi-year upgrade cycles with confidence.
  • Managed service benefits preserved: All RDS managed-service capabilities—automated backups, Multi-AZ high availability, patching, and scaling—apply to 26ai instances, so AI features come without additional operational burden.
  • Regulatory and compliance alignment: Keeping data in-database for AI inference supports data residency and compliance requirements that prohibit sending production data to external services.

How It's Different

  • Versus self-managed Oracle on EC2: RDS handles patching, backups, failover, and the Bedrock integration configuration plumbing, whereas self-managed deployments require customers to manage all of this manually.
  • Versus Aurora PostgreSQL with pgvector: Oracle 26ai combines a mature enterprise RDBMS, native vector search, graph analytics, JSON duality, and direct LLM invocation in a single engine; pgvector provides vector search but lacks native LLM invocation from SQL and Oracle-specific features like Select AI.
  • Versus using a standalone vector database: AI Vector Search stores embeddings co-located with relational data, eliminating synchronization complexity, additional cost, and the latency of cross-service queries required by standalone vector stores.
  • Versus previous Oracle versions on RDS (19c/21c): Oracle 26ai is the first RDS for Oracle version with native Amazon Bedrock integration and Select AI; prior versions have no built-in LLM connectivity or natural language query capability.
  • Versus external RAG architectures: RAG pipelines built entirely in SQL on 26ai avoid the need for external orchestration frameworks (e.g., LangChain, custom Lambda functions), reducing moving parts and simplifying security posture.

When to Prefer It

  • Existing Oracle workloads adding AI: Organizations already running Oracle on RDS 19c or 21c that want to add generative AI or semantic search capabilities without re-platforming or introducing new data stores.
  • Enterprise applications with strict data governance: Use cases where data cannot leave the database perimeter for AI inference due to regulatory, contractual, or security requirements (e.g., healthcare, financial services, government).
  • Business intelligence self-service: Scenarios where business analysts need to query complex relational schemas without writing SQL—Select AI enables natural language querying against existing Oracle data models.
  • Semantic and hybrid search over relational data: Applications requiring similarity search (e.g., product recommendations, document retrieval, fraud pattern matching) that currently rely on a separate vector store alongside Oracle.
  • In-database RAG applications: Teams building LLM-powered applications that need to ground responses in proprietary enterprise data stored in Oracle, where minimizing infrastructure components is a priority.
  • Graph analytics without a graph database: Use cases requiring relationship traversal and graph queries (e.g., supply chain, org hierarchies, network analysis) over data already resident in Oracle.
  • Greenfield Oracle Enterprise deployments: New projects that require a fully managed, AI-capable Oracle database with long-term vendor support and AWS ecosystem integration from day one.

Availability

  • GA Status: Generally Available as of July 7, 2026; Oracle Database 26ai is a Long Term Support Release.
  • Edition: Enterprise Edition only; Standard Edition 2 (SE2) is not supported for Oracle Database 26ai on Linux x86-64.
  • Regions: Available in all commercial AWS Regions and both AWS GovCloud (US) Regions (us-gov-east-1 and us-gov-west-1).
  • Amazon Bedrock Integration: Supported exclusively on Oracle Database 26ai (26.0.0.0); not available on RDS for Oracle 19c or 21c.
  • Upgrade Prerequisites: Upgrading from Oracle 19c non-CDB requires a two-step process: first convert to CDB architecture, then upgrade to 26ai; direct upgrade from 19c or 21c CDB is supported in a single Modify DB Instance operation.
  • Architecture Requirement: Only the multitenant (CDB) architecture is supported; non-CDB creation is not permitted for Oracle 26ai, consistent with Oracle 21c behavior.
  • Pricing: Follows standard RDS for Oracle pricing under License Included or Bring-Your-Own-License (BYOL) models; Bedrock model invocations are billed separately per Amazon Bedrock pricing; Reserved Instance discounts of up to 48% apply.
  • Bedrock Network Requirement: The RDS instance must have outbound HTTPS (port 443) access to the Bedrock runtime endpoint; AWS PrivateLink via a VPC interface endpoint is the recommended configuration.

Tags

Servicesother-aws
Typenew-featurega-launchintegration
Conceptsgenaillmragembeddingnlp
Use Casesenterprisedeveloper-tools
Providersanthropicamazonmeta
GeographyGlobal

Related Resources

AI Radar AWS

AWS AI/ML news — curated, researched, explained

An automated intelligence platform that curates, researches, and analyzes AWS AI/ML/GenAI announcements daily. Every report is backed by real research — the system reads linked blog posts and documentation to provide accurate, in-depth analysis.

How Each Report Is Generated

  1. Collection — Daily monitoring of the AWS "What's New" RSS feed
  2. Filtering — AI-powered relevance detection for AI/ML/GenAI topics
  3. Taxonomy Tagging — LLM-based classification across 6 dimensions
  4. Importance Scoring — Point-based system with tag bonuses (1-5 stars)
  5. Research Phase — Follows links to blog posts and documentation
  6. Report Generation — Claude Sonnet produces structured 6-section analysis
  7. Visual Summary — Claude Opus generates Mermaid diagrams for key items
  8. Publishing — Static website rebuilt and deployed via CloudFront

Features

  • Faceted filtering by service, type, concept, and more
  • Multi-dimensional taxonomy with 80+ tags across 6 dimensions
  • Geographic availability badges (Global, APJ, EMEA, AMER) with filtering
  • Timeline visualization of announcement volume
  • PDF export for offline reading
  • Mermaid visual summaries for key announcements
  • Daily automated updates — no manual curation
What makes this different: Each report involves a dedicated research phase where the system reads linked blog posts and AWS documentation pages. This produces analysis that goes beyond the original announcement text.

Technology

Built with Python, AWS Lambda, Amazon Bedrock (Claude Sonnet 4.6, Opus 4.6, Haiku 4.5), S3, CloudFront, WAF, EventBridge, and CDK.

Open Source

This project is open source. Fork it, customize it for your needs, and deploy your own instance.
📦 github.com/bbonik/ai-radar-aws

How Importance Scoring Works

Each announcement receives a point score based on multiple factors. The total score maps to a 1-5 star rating:

1★ < 2 pts 2★ ≥ 2 pts 3★ ≥ 3.5 pts 4★ ≥ 5 pts 5★ ≥ 6.5 pts

Point Breakdown

FactorPointsWhen
Core AI service (Bedrock, AgentCore, SageMaker AI)+4Service named in title
Key AI service (SageMaker, Kiro, QuickSight)+2Service named in title
Other AI-related service+1Default
Blog post link+3Link to aws.amazon.com/blogs/
GitHub samples link+2Link to github.com/aws*
Documentation link+1Link to docs.aws.amazon.com/
New model+1.5Tagged as "new-model"
New service+1Tagged as "new-service"
New feature+0.5Tagged as "new-feature"
Anthropic / OpenAI provider+2Provider explicitly mentioned
Instance / notebook announcement-2Hardware/capacity, not feature
Performance / pricing / security-0.5Incremental updates
Region expansion to APJ+1Expands to Asia Pacific
Region expansion (non-APJ only)-1.5Only expands to other regions

Geographic Relevance Badges

Each announcement card shows a small badge indicating whether the feature is available in your region:

🌐 Global Available in all regions
🌏 APJ Asia Pacific
🌍 EMEA Europe / Middle East / Africa
🌎 AMER Americas (US, Canada, South America)
No badge Geography unknown
How geography is detected: The system detects ALL geographies mentioned in each announcement. If the text mentions specific regions (Tokyo, Frankfurt, Oregon, etc.), the corresponding geography badges are shown. If it says "all regions" or is a new feature with no region specified, it gets the Global badge. Geography is also filterable — click a geo chip to see only announcements available in that region.