Industry executives and experts share their predictions for 2026. Read them in this 18th annual VMblog.com series exclusive.
By Jags Ramnarayan, Cloud CTO, MariaDB
AI is driving the most significant architectural transformation, databases have seen in decades. The shift from static queries and rigid schemas to dynamic reasoning systems represents a fundamental reimagining of how we interact with data. Natural language interfaces, self-optimizing engines, and intelligent governance are replacing traditional approaches.
The below predictions map out how AI will reshape every touchpoint, from query writing to insight generation – transforming how developers, analysts, and enterprises leverage their data.
The Shift from API Calls to Agentic Intelligence
The next decade will see applications evolve from simply calling APIs to becoming increasingly agentic – capable of planning multiple steps ahead, testing hypotheses, and providing recommendations with reasoning. This fundamental shift will transform how databases operate, moving from neat OLTP lookups to complex explorations that require real-time answers without cobbling together multiple products.
Natural Language Will Dominate Database Interactions
SQL won’t disappear, but it will become an artifact rather than the primary interface. Developers, analysts, and operators will interact with databases through natural language, with platforms automatically translating requests into SQL and providing explanations. Every database platform will need embedded semantic layers that understand schemas, relationships, and business terminology, plus planning capabilities to decompose complex requests into executable steps.
“Instant RAG” Will Become Table Stakes
RAG (Retrieval-Augmented Generation) capabilities will be built directly into database platforms rather than requiring separate systems. Platforms will natively ingest documents, embed and index them, and make them joinable with traditional row data. This convergence means a single query can seamlessly touch both documents and tables, returning answers with citations and confidence scores.
The OLTP/Analytics Boundary Will Dissolve
The distinction between transactional and analytical workloads will blur as agentic patterns demand both capabilities simultaneously. Future platforms will deliver “Operational Analytics with Transactions” through architectures featuring one SQL surface with multiple engines (transactions, analytics, vectors) that scale independently while maintaining data freshness within seconds.
AI-Native Development Will Be the Default
Development teams will build with AI from day zero, even if they don’t immediately ship agent-based features. Database-as-a-Service platforms will be conversational by default, with natural language prompts generating SQL, migrations, and optimizations. Copilots will analyze query plans, suggest indexes, and surface performance issues with automated remediation steps.
Governance Will Be Embedded in Every Response
As conversational interfaces become standard, governance and explainability will travel with every answer. Each natural language result will include the underlying SQL, document citations, vector distances, consistency classes used, and confidence scores. Audit trails, cost/time budgets, and human-in-the-loop gates for high-stakes actions will be built into the conversation flow.
The “Push-Button” Era of Data Platform Capabilities
Complex capabilities that currently require extensive engineering will become push-button features. RAG implementations, multi-engine orchestration, and AI-powered optimizations will be available out-of-the-box rather than requiring months of custom development.
The convergence of cloud and AI isn’t just changing how we build databases, it’s fundamentally reimagining the entire developer and analyst experience. The winners in this space will be platforms that deliver unified surfaces with multiple specialized engines, conversational interfaces with SQL transparency, and operational discipline at scale. The future isn’t about replacing human judgment but dramatically shortening the path from intent to insight while maintaining enterprise-grade governance and reliability.
##
Jags is a data technology veteran and the Cloud CTO for MariaDB plc. With over 20 years leading distributed data products, his career includes co-founding VMWare GemFire (now Apache Geode), a pioneering distributed in-memory data grid. Later, he co-founded SnappyData (acquired by Tibco) to define a new unified data fabric, where he helped lead work on hybrid translytical processing. Most recently, he leveraged his cloud-native expertise to help develop the world-class SkySQL database-as-a-service, which was acquired by MariaDB.
