OpenAI announced plans to acquire Ona as part of an effort to strengthen Codex with secure cloud execution and orchestration capabilities. The company said Codex has grown rapidly, reaching more than 5 million weekly users, a figure that represents a 400% increase compared with earlier this year. Originally focused on software development, Codex now supports a broader set of tasks, helping people research, analyze, create, and automate work from an initial request through completion.

The acquisition centers on extending how Codex operates. According to OpenAI, increasingly valuable tasks handled by Codex now span hours or even days rather than brief sessions. The company aims to make it possible for work to continue independently of the device where it started. Users would be able to monitor progress, provide guidance, make decisions, and review outcomes while agents continue operating in the background. Ona’s technology is designed to provide the persistent and secure environments needed for that model.

Ona brings experience in moving software development workflows from local systems into cloud-based environments. The company says it has helped 2 million developers work within secure and reproducible cloud setups and already serves several customers shared with OpenAI. That background aligns with OpenAI’s stated objective of enabling agents to continue working within customer cloud environments even when users are no longer actively connected through their devices.

The announcement also highlights a broader shift from experimentation to operational deployment. OpenAI said organizations adopting agents across production workflows require more than advanced models. They need confidence that systems meet security, governance, and operational standards. The company emphasized requirements such as control over execution environments, access permissions, credential management, activity monitoring, and review processes. Ona’s customer-controlled execution approach is intended to address those needs by allowing agents to operate within an organization’s own cloud infrastructure.

The acquisition remains subject to customary closing conditions and regulatory approvals. Until the transaction closes, OpenAI and Ona will continue operating independently. Following completion, the Ona team is expected to join OpenAI and collaborate with the Codex team on secure enterprise execution capabilities. OpenAI said the combined effort will support engineering teams handling sustained work across software development activities, including testing, issue resolution, application modernization, vulnerability remediation, and other complex workflows that unfold over time.

📊 What This Means (Our Analysis)

This development stands out because it focuses on the environment in which agents operate rather than solely on model capability. OpenAI’s announcement suggests that long-duration work is becoming an increasingly important use case for Codex, creating demand for infrastructure that allows tasks to continue beyond a single session. By integrating persistent execution technology, the company is positioning Codex to support work that unfolds over extended periods while maintaining continuity and oversight.

The emphasis on customer-controlled environments is equally notable. Throughout the announcement, OpenAI repeatedly highlights security, governance, operational controls, and infrastructure ownership. That focus indicates that broader enterprise adoption depends not only on agent performance but also on giving organizations confidence in how those agents run. The planned combination of Ona’s execution model and Codex’s orchestration capabilities aims to address that requirement while expanding the practical scope of agent-driven work.

📌 Our Take: The next phase of Codex appears focused on turning extended, secure agent workflows into a standard part of enterprise operations.