OpenAI to Acquire Ona, Pushing Codex Toward Persistent Cloud Agents

OpenAI will acquire Ona, folding the company's secure cloud execution and orchestration technology into the Codex ecosystem, the company announced today.

The acquisition targets a specific structural problem in agentic AI: work that no longer fits inside a single session. More than 5 million people now use Codex each week to research, analyze, build, and automate their work — a figure OpenAI says is up 400% from earlier this year. Codex began as a tool for software developers and now serves a wider range of users taking complex tasks from an initial request through to a finished result.

As Codex becomes more capable, its most valuable work unfolds over hours or days rather than minutes, according to OpenAI's announcement. "We believe people should be able to delegate more ambitious work without remaining tied to the machine where it began," the company said. "The work should continue beyond the initial session, with Codex making it possible to stay connected and check progress, provide direction, make decisions, and review results from anywhere."

Ona's technology provides the missing infrastructure: secure, persistent environments where agents can access the tools, systems, and context they need to make progress over time. OpenAI frames the deal as a way to expand Codex beyond work tied to a single device or active session, and to help more organizations deploy agents securely in production.

Why persistence matters

Ona has spent years moving software development from local machines into the cloud. The company says it has helped 2 million developers work in secure, reproducible cloud environments and already supports multiple shared customers with OpenAI.

That track record maps directly onto the next phase of Codex. Ona's environments allow agents to continue working inside a customer's cloud environment even when laptops are closed — the difference between an assistant that stops when you walk away and one that keeps a multi-day refactoring or migration task moving overnight.

The distinction matters commercially. Enterprise buyers evaluating agent platforms increasingly distinguish between chat interfaces and systems capable of sustained autonomous work. Persistent execution — with continuity of state, tool access, and context across sessions — is a precondition for the latter, and it is infrastructure that model capabilities alone do not provide.

The enterprise control question

The acquisition also speaks to the governance concerns that dominate enterprise AI adoption. As organizations move from experimenting with agents to deploying them across production workflows, capable models are only one part of what they need, OpenAI notes.

"Organizations should be able to give people the benefits of persistent agentic work with confidence, knowing agents are operating inside environments that meet their security, governance, and operational requirements," the announcement states. "That means having control over where they run, what they can access, how credentials are scoped, how activity is logged, and how work moves through review."

Ona's customer-controlled execution model addresses this directly. Agents will operate inside an organization's own cloud environment while OpenAI provides the intelligence and orchestration that power the experience. The arrangement gives organizations control over their infrastructure, data, and security boundaries without limiting what Codex can accomplish, according to the company.

That positioning puts OpenAI in direct competition with infrastructure-led approaches to agent deployment, where the pitch rests on keeping execution inside the customer's perimeter. Buying Ona rather than building the capability in-house shortens OpenAI's path to that enterprise segment — though the purchase price and terms were not disclosed in the announcement.

Deal terms and next steps

The acquisition remains subject to customary closing conditions, including receipt of required regulatory approvals. Until closing, OpenAI and Ona will remain separate and independent companies.

After closing, the Ona team will join OpenAI and work with the Codex team to advance secure, persistent enterprise execution capabilities and scale Codex to more enterprises worldwide.

OpenAI says the combined effort will help engineering teams take on sustained work across the software lifecycle — from running tests and resolving issues to modernizing applications, addressing vulnerabilities, and supporting complex workflows over time.

"We're excited to welcome the Ona team to OpenAI as we build AI systems that help people and businesses create and accomplish more while meeting the security and control requirements of real-world deployment," the company said.

With weekly Codex usage up 400% since early this year and workloads stretching from minutes into days, OpenAI is betting that the next competitive frontier in agentic AI is not model intelligence but where and how agents run. If the deal closes as expected, Codex gains a persistent place to work — and OpenAI gains a stronger foothold in the production environments where enterprise agent deployments will be won or lost.