MariaDB Builds Its Agentic Enterprise Platform Around Live Data

MariaDB wants agents on one foundation. MariaDB Enterprise Platform 2027 brings 56% higher peak throughput, stronger resilience, and new AI capabilities to a durable platform built around live operational data. The company also introduced a new brand identity to mark its shift from database vendor to platform provider for the agentic enterprise.

That shift reflects a practical problem: enterprises already depend on transactional systems, and most will not rebuild their data infrastructure to adopt agents. IDC predicts that by 2027, G2000 companies—the world’s 2,000 largest companies—will increase agent use tenfold, while token and API call loads rise a thousandfold.

IDC also predicts that companies without high-quality, AI-ready data foundations will suffer a 15% productivity loss by 2027. MariaDB’s answer is to add agents to the systems enterprises already run instead of bolting on another brittle layer. Every extra system adds a seam, and seams break. Technology has discovered this several times, apparently without learning from it.

One platform for transactions, analytics, and agents

MariaDB Enterprise Platform 2027 builds outward from a transactional core already running mission-critical systems across financial services, telecommunications, retail, and the public sector. What began as an open source operational database now covers mission-critical transactions, real-time analytics, in-memory speed, AI tooling, and multi-region scale on one unified platform.

The platform keeps its code open while MariaDB verifies contributors, vets releases, and ships hardened builds. That gives enterprises the transparency of open source with a vendor accountable for the resulting platform, including deployments on premises, in any cloud, or as a managed service on MariaDB Cloud.

Rohit de Souza, CEO of MariaDB plc, said, “The AI race won’t be won by the smartest model. It will be won by the companies that unify the infrastructure underneath their agents.” He said agents need a durable foundation that does not shift beneath them, allowing organizations to trace every action they take.

The platform’s AI features target the infrastructure around an agent rather than only the model itself. AI Retrieval grounds agents in documents and live operational data through one managed service, turning natural-language questions into read-only SQL. AI plugins bring MariaDB expertise into coding agents such as Claude Code and Codex, while Agent Memory gives long-running agents a durable in-memory store shared across sessions.

Semantic Cache returns stored answers when users ask an existing question in different words, cutting response times and token costs. MariaDB Exa runs real-time analytics on live operational data without complex pipelines, and native vector search delivers up to 30% more queries per second than the last major release.

Performance, governance, and a new identity

Third-party testing puts peak throughput 56% above the last major release. Raft clustering, conflict detection, and conflict resolution add scale and resilience for multi-region deployments, while one set of role-based access controls spans every workload.

Bring Your Own Cloud lets enterprises keep data where they choose, and FIPS 140-3 supports cryptography requirements for federal and other regulated enterprises. The platform includes MariaDB Enterprise Server 12.3, MariaDB Exa, MariaDB GridGain, MariaDB MaxScale, MariaDB Enterprise Manager, and the MariaDB Enterprise Kubernetes Operator.

Devin Pratt, Research Director, Database Systems for Enterprise Intelligence at IDC, said, “The database has quietly become the hardest problem in enterprise AI.” His practical conclusion is less glamorous than another model announcement: enterprises will add agents to the systems that already run the business, making the transactional core a starting point for analytics, retrieval, and memory from the same live data.

Sanjeev Mohan, Principal Analyst at SanjMo, made a similar case. “Most platforms are racing toward agents from the analytics side, which means working backward into transactions, the hardest workload to get right,” he said, arguing that building outward from a proven transactional core is the more durable path.

That foundation still needs boundaries. Even when agents perform reliably in testing, certain actions warrant human oversight, and technology leaders must enforce limits through system architecture rather than prompts or trust alone.

A semantic-integrity gate can verify that each record’s meaning matches its metadata before ingestion—not only its schema. An invoice date can be valid yet mislabeled as a payment date; quarantining that mismatch before a model or vector store sees it prevents an answer grounded perfectly in the wrong business meaning.

MariaDB Enterprise Platform 2027 is the culmination of two years of strategic investment and engineering. It is generally available today, and MariaDB AI Retrieval follows in public preview on MariaDB Cloud later this month.

The release launches alongside a redesigned MariaDB website and new logos and brand guidelines. Tracy Strauss, CMO of MariaDB plc, said, “A brand should tell you what a company believes before you read a word, and ours was telling an old story.” The new identity is intended to show a company that is open by design, unified across workloads, and built for the agentic enterprise.

A webinar titled “Built for the Agentic Era: Introducing MariaDB Enterprise Platform 2027” is scheduled for November 18, 2026. The pitch is clear: run transactions, analytics, and AI on the same live data instead of stitching systems together—and hope the seams hold.

Stay connected via Google News

Follow us for the latest AI News