日本語 
Oracle just made a very clear bet. Stop moving data around. Bring AI to where the data already sits.
The company has introduced agent-based AI capabilities inside its AI Database, aimed at helping enterprises build and run AI agents on real business data without the usual mess of pipelines and integrations. The core idea is simple but important. Instead of stitching together models, tools, and databases, Oracle is trying to collapse everything into one system that handles data access, reasoning, and execution.
What stands out is the focus on control and security. Features like role-based data access, private AI model deployment, and ‘trusted answer’ systems are clearly a response to growing concerns around data leaks, hallucinations, and prompt injection. This is not just about smarter AI. It is about safer AI that enterprises can actually trust in production.
Also Read: New Relic Set to Open Its First Data Center in Japan
The situation has a second aspect which shows. The system enables multiple cloud environments and on-site installations together with open data format support which helps organizations avoid becoming dependent on any particular vendor.
Big picture. This reflects where the market is heading. AI is moving from experiments to core systems. And whoever owns the data layer is trying to own the AI layer too.
English