Earlier this year, we launched OCI Enterprise AI to help organizations simplify how they build, deploy, and scale AI solutions. Since then, we’ve continued to expand its capabilities to help businesses take advantage of the latest AI innovations while reducing the complexity of putting AI to work.

As organizations build more sophisticated AI applications and autonomous agents, two challenges are becoming increasingly important: ensuring reliable access to AI models across regions and providing secure environments where agents can execute tasks.

Today, we’re introducing two new capabilities in OCI Enterprise AI—Smart Model Router and Agent Sandbox—to help organizations address these challenges and scale agentic AI globally with greater confidence.

Expanding AI access across 30 additional OCI regions with Smart Model Router

AI model availability and GPU capacity can vary across cloud regions, making it challenging to deploy AI applications globally. Developers often need to manage regional model endpoints and build custom routing logic, adding complexity and operational overhead.

To address these challenges, Oracle is expanding OCI Enterprise AI to 30 additional commercial regions, bringing supported on-demand models and selected Enterprise AI features, including hosted deployments, sandboxes, and API keys, to more locations where customers operate.

Smart Model Router enables this expanded model access by routing inference requests to GPU-backed regions, without requiring the selected models to be hosted locally. Customers define the supported model and permitted destination regions through reusable routing profiles, while OCI automatically routes inference requests to approved regions with available capacity.

For businesses, this means greater flexibility to deploy AI applications globally, improved availability, and less time spent managing regional infrastructure—all while maintaining control over where inference is processed.

To learn more about how it works, see OCI Enterprise AI Smart Model Router and OCI Enterprise AI new supported regions. You can also learn more with the model routing tutorial.

Enabling secure agent execution with Agent Sandbox

As AI agents take on increasingly complex tasks, businesses need a way to give them greater autonomy without compromising security and control.

OCI Enterprise AI Agent Sandbox is now generally available, offering a managed execution layer for agentic applications. An agent can create temporary Linux workspaces to test generated code and run shell commands, analyze business data, or run diagnostic scripts, then collect the results and stop the workspaces.

Organizations configure compute resources, execution duration, and network access while OCI operates the underlying infrastructure. This gives teams a controlled place to execute agent tasks without building and maintaining a separate execution platform.

Agent Sandbox helps organizations adopt more capable AI agents while reducing operational overhead and maintaining enterprise security controls. Learn more about Sandboxes in OCI Enterprise AI.

What can businesses build with these capabilities?

From global AI applications to intelligent workflow automation, these capabilities open up practical opportunities across industries.

  • Global AI applications: Extend AI-powered customer and employee experiences to more regions, accessing model capacity in approved OCI locations without requiring local GPU infrastructure.
  • Intelligent business analytics: Enable AI agents to analyze data, execute code, and generate reports or visualizations in secure environments.
  • Agentic workflow automation: Build agents that process files, use approved tools, and carry out multi-step business tasks within controlled execution boundaries.

Expanding the foundation for scalable agentic AI

Smart Model Router and Agent Sandbox address two essential requirements for scaling agentic AI: reliable access to models and controlled execution of agent-driven tasks.

Smart Model Router and Agent Sandbox address two essential requirements for scaling agentic AI: reliable access to models and controlled execution of agent-driven tasks.

Combined with expanded regional availability, these capabilities make it easier for organizations to build global AI applications, automate business workflows, and deploy more capable AI agents. By simplifying infrastructure management and strengthening operational control, OCI Enterprise AI helps businesses focus on what matters most—turning AI innovation into business value.

Learn more about OCI Enterprise AI: