Couchbase plots course to be broader AI data platform

Couchbase launched an AI data layer and other enhancements to expand beyond a database provider to be an AI data platform. Couchbase's plan is to focus on mission-critical applications, infrastructure consolidation and AI-ready data.

The company's summer release cycle aims to reposition Couchbase to take on the likes of Oracle, MongoDB and other database companies that operate as platforms. The new Couchbase, courtesy of a revamped product roadmap and influx of talent, revolves around mission critical applications, a unified platform to cover operational data and AI insight, and multiple deployment options beyond the cloud.

Couchbase was formed via the 2011 merger of Membase and CouchOne and completed an IPO in 2021. In 2023 through 2025, Couchbase shifted to focus on cloud services and expanded enterprise features. Couchbase went private in September 2025 and for the six months of 2025 had revenue of $114.9 million.

In addition, Couchbase has a revamped leadership team that features a bevy of AWS, Oracle and Google Cloud alums including Barry Morris, Chief Product and Strategy Officer, and Tim Finley, SVP of AI. In a briefing with analysts, Morris and Finley walked through Couchbase's summer releases and strategy.

Here's a look at the architecture Couchbase is betting on.

The company's big pitch revolves around total cost of ownership. Couchbase claims that 53% of customers have realized more than a 50% TCO gain from running the most critical applications compared to the systems they replaced.

"This is a system that's a distributed system first and a database system second. Everything that you see here, you should think of as microservices. These are all independent services that run in a distributed way, which scale out independently from each other. It genuinely is a single unified system. It is not a bag of checkbox kind of features that we've added to a monolithic system," said Morris.

Couchbase's AI Data Plane is the headliner of the companies summer launches. AI Data Plane is designed to address the bottleneck in AI, which is the data layer. Couchbase AI Data Plane is one brand for managed and self-managed deployments. The self-managed version of AI Data Plan ships components including Agent Memory, MCP Server and Agent Catalog.

AI Data Plane includes the following:

• Enterprise Analytics 2.2, which includes Apache Iceberg federation, Google Cloud Storage as a native layer and Oracle and SQL Server hooks, Couchbase's Capella Analytics is now available on Azure.
• Agent memory in AI Data Plane has short- and long-term memory, access control and validation for LangGraph, CrewAI and Llama Index.
• A set of flexible tools for building agents with choices of models, embeddings on demand, open agent access and visibility.

Morris added:

"The AI data plane is really about delivering AI ready data in real time to tens of thousands of agents. That's the leadership position that we're aiming at."

Finley emphasized that Couchbase's AI services in its Capella managed services are available. New functionality will enable support to customers that decide to run AI on their own infrastructure.

"We need to support customers running AI everywhere, which is why you are going to see from us continual investment in AI data plane, which is going to offer customers AI functionality, whether they're running it in the cloud or on premise, or whether they're managing it or we are managing it," said Finley. "Agents are going to be the primary way we interact with computers of the future. Doing AI in an enterprise requires a lot more than models, and it ends up coming down to the data, more importantly, not just the data, but how you are putting the data to work."

Couchbase also rolled out the following to add to its platform.

• Sync Gateway in Couchbase's core platform supports rolling upgrades, enabling zero-downtime maintenance end to end and not just for the core database. Sync Gateway also has distributed resync for large datasets and metadata isolation improvements.
• Edge Server is a Couchbase variant for constrained environments and edge computing.
• Capella enhancements include private link support for live migrations from self-managed to Capella.
• Enterprise Analytics has a column store for JSON for operational analytics with new features for automated index recommendations, index-only query plans and Rust SDK.
• Cloud catalog integrations are native for AWS Glue, Databricks Unity Catalog and Google Cloud's native catalog.

Whether Couchbase is successful will depend on its overarching strategy in an increasingly crowded space. Morris said the new Couchbase is dramatically different from what it was before it went private. Key points about the strategy:

• Couchbase is focused on the global 2000 enterprises.
• The company will focus on "world champion" use cases that can be differentiated with mission-critical scale and AI.
• Couchbase will focus on fewer, higher-value segments and emphasize customer success and co-invention.
• The company is planning to enable customers to use TCO savings from consolidation to fund AI initiatives.