OpenAI has agreed to acquire Ona, a cloud infrastructure startup focused on secure development environments, as it expands Codex from a coding assistant into a platform for long-running AI agents operating inside enterprise systems.
The acquisition targets a key challenge facing enterprise AI deployments, which is to enable agents to continue working after a user leaves a session. This also allows maintaining control over security, credentials, governance and infrastructure.
OpenAI said Ona’s technology will allow Codex agents to run within customer-controlled cloud environments, giving organisations greater oversight of where agents execute, which systems they can access, and how their activity is reviewed.
While AI agents have become increasingly capable of handling complex tasks, many remain constrained by session-based interactions and limited access to enterprise environments.
OpenAI said more than 5 million people now use Codex each week for software development, research, analysis and workflow automation, a figure that has grown 400% since earlier this year.
According to the company, some of the most valuable work performed by Codex now unfolds over hours or days rather than minutes, increasing demand for persistent execution environments that remain available beyond a single device or active session.
Founded to help developers move software development workloads from local machines into cloud-based environments, Ona says its platform has been used by more than 2 million developers.
Ona provides cloud-based execution environments designed specifically for AI software agents. Rather than running on a developer’s laptop, agents operate inside isolated cloud workspaces with access to code repositories, development tools, network resources and enterprise systems.
The platform also includes governance features such as credential management, audit logs, security policies and customer-controlled access controls, allowing organisations to run agents within their own cloud infrastructure.
OpenAI plans to integrate that technology into Codex, allowing agents to continue operating within an organisation’s own cloud infrastructure while OpenAI provides the models and orchestration layer.
Calling it the “next phase of Codex”, OpenAI says that this integration will enable agents to “continue working inside a customer’s cloud environment even when laptops are closed.”
The company said the approach is designed to meet enterprise requirements around data residency, access controls, credential management, logging and review processes.
“Agents need more than intelligence; they need a trusted workspace,” said Johannes Landgraf, Ona’s Co-Founder and Chief Executive. “We built Ona to give agents cloud environments with the context, control and collaboration enterprises require.”
Thibault Sottiaux, OpenAI’s Core Products Lead, said enterprises increasingly want AI agents that can perform operational work while meeting existing security and compliance requirements. “Ona will help us make Codex easier to deploy securely across production workflows for customers operating at the highest standards of trust and scale," he said.
The acquisition remains subject to customary closing conditions and regulatory approvals. OpenAI said the Ona team will join the company after the transaction closes and will work on enterprise execution capabilities for Codex, including software engineering workflows such as testing, issue resolution, application modernisation, and vulnerability remediation, which may require agents to operate continuously over extended periods.
The Ona acquisition adds to a series of OpenAI transactions aimed at expanding Codex beyond its core models. In March, OpenAI agreed to acquire Python tooling company Astral, bringing its development tools into the Codex platform as competition intensified with Anthropic's Claude Code. OpenAI said at the time that Codex had surpassed 2 million weekly active users.
Your reaction
Discussion
What's your take on this story?
No comments yet — be the first to share your view.