OpenAI just bought the infrastructure to keep AI agents alive for days. Your enterprise deployment just got harder.

OpenAI acquired Ona — formerly Gitpod — to give Codex agents persistent cloud environments that run for hours or not just seconds.

This is the infrastructure layer nobody talks about.

Your enterprise AI agents die when context is lost between tool calls. Ona keeps them alive with sandboxed machines pre-configured with your codebase, tools, and access controls.

Codex now has 5 million weekly users. Up 65% in two months. That scale demands execution infrastructure that doesn't break isolation between enterprise tenants.

OpenAI is now competing directly with Salesforce Agentforce and ServiceNow Action Fabric. Both announced similar infrastructure investments in May.

If you're piloting Codex, your procurement picture just changed. Ona replaces the workaround of spinning up VMs manually and managing state externally.

Audit your agent deployment architecture. The question is no longer which model. It's which execution layer keeps your agents alive without compromising security.