MongoDB Launched Version 9.0 and Atlas AI Engine
The latest release delivers significant query performance gains and new tooling for AI agent memory and governance.
Updated on Sept. 29, 2026 in Artificial Intelligence
Live Poll
Do you trust integrated vendor platforms more than custom-built frameworks for your organization's AI agents?
MongoDB has released version 9.0 of its database alongside the new Atlas AI Engine, a suite of tools designed to simplify the deployment of AI agents. The updates also introduce Atlas Infinite, a deployment option currently in public preview on AWS that disaggregates compute and storage for independent scaling.
Why it matters
The release enables organizations to handle the volatile, high-traffic demands of AI agents while improving overall database efficiency. By separating storage and compute, Atlas Infinite aims to provide the elasticity required for modern transactional workloads that scale dynamically.
MongoDB 9.0 delivers a 2x increase in throughput on large instances and 20 percent higher throughput for transactional workloads compared to version 8.0. Additionally, the platform now supports search functionality directly on encrypted data.
The players
MongoDB
A database company that develops document-oriented storage systems and the Atlas cloud-native platform.
Voyage AI
An AI research firm specializing in embedding models and retrieval systems, acquired by MongoDB in 2025.
The details
Atlas Infinite works by disaggregating compute from storage, allowing both resources to scale independently to accommodate AI agent traffic spikes. The Atlas AI Engine leverages embeddings from Voyage AI, a firm acquired in 2025, to provide native retrieval capabilities that serve as memory for AI agents. This architecture simplifies the stack by centralizing identity, governance, and data retrieval, which previously required separate, disparate services.
Timeline
MongoDB acquired Voyage AI in 2025.
MongoDB 9.0 and the Atlas AI Engine launched on September 29, 2026.
The Tech Race
The introduction of the Atlas AI Engine places MongoDB in direct competition with emerging specialized vector databases and agent-orchestration platforms. By integrating native retrieval and memory management, the company aims to retain enterprise workloads that might otherwise migrate to separate AI-specific infrastructure providers.
Developers can now access the Atlas AI Engine in public preview, while Atlas Infinite is available for AWS users as a new deployment option. Users should note that the full feature set, particularly for government-specific versions, remains in the planned release roadmap.
The takeaway
MongoDB is betting that shifting the burden of AI memory and governance into the database layer will define the next generation of application development. Watch for the official release of Atlas Infinite for Government and the expansion of the platform to additional cloud providers.
Further reading
For broader trends in AI-native data management, visit the Artificial Intelligence section.
More information
For setup and deployment instructions, visit the MongoDB Atlas registration and start page.
Live Poll
Do you trust integrated vendor platforms more than custom-built frameworks for your organization's AI agents?
