The open-source MariaDB database now has a vector search capability, allowing customers’ businesses to access a single, complete database supporting transactional, analytical, semi-structured, and now artificial intelligence (AI) applications.
The firm has undergone a private equity makeover since going public in 2022 with a $672 million valuation. It posted poor quarterly results, received NYSE warnings about its low market capitalization, laid off 28 percent of its staff, closed down its MariaDB Xpand and MariaDB SkySQL offerings, and announced Azure Database for MariaDB would shut down in September 2025. It was a financial storm, with one lender even threatening to sweep its accounts. It received many takeover proposals and accepted one from K1 Investment Management valuing it at $37 million.
Thus MariaDB, with its nearly 700 active customers across industries including banking, telecommunications, government, healthcare and e-commerce, went private in September 2024. At the time, Sujit Banerjee, MD of K1 Operations, said: “Together, we aim to accelerate product innovation and continue MariaDB’s mission of delivering high-quality, enterprise-grade solutions to meet the growing demands of the market.” K1 appointed Rohit de Souza as CEO while former CEO Paul O’Brien remained involved as an advisor.
An exec replenishment exercise included MariaDB rehiring Michael “Monty” Widenius as CTO. He is the creator of MySQL and previously had a 7 year stint as Maria CTO up until December 2022. The firm also appointed Vikas Mathur as Chief Product Officer and Mike Mooney as Chief Revenue Officer. It said it would launch vector search in MariaDB Server and a Kubernetes (K8s) Operator, catering to AI and cloud-native trends, and it has now done so.
The MariaDB Enterprise Platform 2025 version delivers:
- Native vector search capabilities within the core database engine that is 100 percent open source
- Additional JSON functionality for expanded semi-structured data support
- A new version of MaxScale for worry-free upgrades
- A new cost-based optimizer to support complex queries

Vikas Mathur stated: “Bringing vector search natively to the database server lets customers extend the same database they’re already using throughout their organization to new AI applications. We’ve also prioritized peace of mind for our enterprise customers with this release, adding tools to make upgrading versions a breeze.”
It says the vector search enables businesses to keep their database stack simple while supporting AI initiatives. MariaDB says vector search enables searching unstructured data by value and by semantics without the need to integrate multiple databases or compromising your system’s reliability or security. It can also help LLMs deliver more accurate and contextually relevant results using retrieval-augmented generation (RAG) on enterprise data.
Also, by leveraging their existing MariaDB database, customers can eliminate the need to maintain separate vector databases.
Benchmarks of MariaDB’s vector search capabilities show that it outperforms PostgreSQL’s pgvector by 1.5x for queries per second (QPS) and 2x on index creation time.
The new MaxScale feature allows customers to capture the workload of a production system, such as queries, sessions and transactions, and replay them in a test environment. The captured workloads can be used to verify that upgrades of the MariaDB database behave as expected and to measure the effects configuration changes may have.
The revamped optimizer in the database engine ncorporates a granular and refined cost model that takes into account state-of-the-art SSD disks and different characteristics of storage engines. MariaDB’s database can now “fully leverage the lower latency and high throughput offered by modern storage devices, automatically choosing the fastest execution plan for complex queries.”
The new release also adds:
- Reduced operational downtime with an online schema change feature that allows for a non-locking ALTER TABLE built into the server, which enables writing to a table while the ALTER TABLE is running.
- Improved efficiency with a new optimistic ALTER TABLE for replication that greatly reduces replication lag by making ALTER TABLE operations two phased operations.
- Enriched enterprise-class protection with new security features that enable TLS encryption by default, add more granular privileges and introduce a new plugin to prevent reuse of old passwords.
- Enhanced partitioning management with new operations like converting partitions to tables and vice versa, and new enhancements to managing system versioning partitions, making database maintenance more flexible and efficient.
MariaDB Enterprise Platform 2025 includes updated versions of its core database, MariaDB Enterprise Server 11.4, its advanced database proxy, MariaDB MaxScale 25.01, and tools and support. The MariaDB Enterprise Platform is available to download now for all MariaDB customers.
For more information read the launch blogs: Announcing New Release of MariaDB Enterprise Platform, Introducing Vector Search with the Latest Version of MariaDB Enterprise Platform and Introducing MariaDB MaxScale 25.01 GA.