OpenAI Acquires Ona to Enhance Secure AI Agent Capabilities

OpenAI is acquiring Ona to integrate secure, persistent cloud environments into its AI agent infrastructure. According to statements from OpenAI and Ona CEO Johannes Landgraf, the acquisition allows AI agents to maintain context and access tools across multiple sessions and devices, enabling enterprises to deploy agents securely in production.

Why did OpenAI acquire Ona?

OpenAI intends to move its AI capabilities beyond single-device tasks or active sessions. An official OpenAI statement confirmed the company will use Ona’s technology to expand Codex, providing agents with “secure, persistent environments” where they can access the tools and systems required to make progress over time.

This shift addresses a primary hurdle for corporate AI: the “memory” gap. Most AI interactions are ephemeral. By integrating Ona, OpenAI creates a space where an agent’s work continues even after a user closes a browser tab or switches devices.

How does this impact enterprise AI deployment?

The acquisition targets the specific security requirements of highly regulated industries. Johannes Landgraf, CEO of Ona, stated that the platform provides “trusted, customer-controlled cloud environments” where software actually lives.

Landgraf reported that weekly Ona agent sessions grew 13x since the start of the year. He attributed this growth to adoption by “demanding institutions,” specifically citing the oldest bank in the US, one of Europe’s largest pharma companies, and one of Asia’s largest sovereign wealth funds.

The shift from chatbots to production agents

Most companies currently use AI as a chatbot—a tool for drafting emails or summarizing text. The Ona integration signals a move toward “production agents.” These are autonomous entities that operate inside a company’s own systems to execute multi-step workflows over days or weeks.

What happens next for AI agents in the workplace?

The integration of frontier intelligence with persistent environments suggests a future where AI agents act as permanent digital employees. According to Landgraf, OpenAI provides the “frontier intelligence” and “scale of research,” while Ona provides the “building blocks” for enterprise work.

This combination likely leads to agents that can manage long-term projects—such as auditing a year’s worth of financial records or monitoring a drug trial—without requiring a human to keep the session active.

Industry analysts often contrast this approach with “stateless” AI. While stateless AI treats every query as a new beginning, the OpenAI-Ona model treats the AI as a resident of the corporate cloud, possessing a continuous history of its actions and access to live systems.

Frequently Asked Questions

To learn more about how autonomous agents are changing corporate workflows, read our guide on Enterprise AI Implementation or visit the official OpenAI blog.