OpenAI has agreed to acquire Ona, a cloud development environment provider with 79 employees that was formerly known as Gitpod, in a move aimed at making agentic AI more suitable for enterprise use. The acquisition will give Codex users self-hosted sandboxes, allowing tool execution to run in infrastructure that customers control instead of OpenAI’s own services.
CIOs and CISOs have many concerns about deploying fully autonomous agents. Questions include whether an agent might delete critical files, run off on unintended tasks and generate massive token bills, or be manipulated by external actors into performing malicious actions. Ona’s technology is designed to reduce those risks by providing secure, persistent environments where agents can access tools, systems, and context over time.
What Ona Brings to OpenAI
In an OpenAI statement, the company said Ona’s platform offers trusted, customer-controlled cloud environments where work continues across devices and inside the systems where software actually lives. By integrating Ona, OpenAI plans to expand Codex beyond work tied to a single device or active session and help more organizations deploy agents securely in production.
Ona CEO Johannes Landgraf echoed that view, noting that Ona provides the building blocks agents need for enterprise work, while OpenAI brings frontier intelligence, product polish, and a scale of research and distribution that Ona could not reach alone.
Landgraf did not disclose revenue figures, but he hinted at large customers. Since the start of the year, weekly Ona agent sessions have grown 13× in production across institutions including the oldest bank in the U.S., one of Europe’s largest pharma companies, and one of Asia’s largest sovereign wealth funds.
Acquisition Price and Market Context
IDC estimates Ona’s 2025 annual revenue at roughly $7 million, with 2026 revenue likely between $10 million and $15 million. Using a standard 30× revenue multiple for acquisitions, that suggests a price of around $450–$500 million.
IDC sees this as a strategically smart move for OpenAI. While OpenAI has invested heavily in Codex, it lacked a safe, persistent area for enterprise autonomous agents. Ona provides secure environments where agents can have memory and operate under customer control—technology that is likely necessary for enterprise deployments, even if its exact quality is hard to judge.
Competitive Pressure from Anthropic
Gartner’s First Take notes that the acquisition gives Codex the scaling capability it previously lacked, but it also forces enterprises to weigh the benefits of a vendor-specific integrated stack against the flexibility of staying vendor-agnostic. Gartner also says this move appears to be OpenAI’s response to Anthropic supporting self-hosted sandboxes in Claude Managed Agents starting May 2026.
Tom Findling, CEO of Conifers.ai, sees competition with Anthropic as a major factor. He认为 this is less about removing a small competitor and more about ensuring Codex is enterprise-ready before Anthropic gains too much momentum. In the enterprise market, the battle is not just about who has the smartest coding model, but who can make AI agents safe and useful enough for large companies to deploy.
Jason Andersen of Moor Insights & Strategy agrees that OpenAI and Codex have given ground to Anthropic and Claude Code. He emphasizes that the real issue is not making Codex better at writing code, but making it work inside real enterprise environments where security, access controls, persistent cloud workspaces, audit trails, and integration with existing developer workflows are critical. Ona provides some of that missing plumbing.
Why This Matters for Enterprise Adoption
Jeremy Roberts of Info-Tech Research Group sees this as a positive step for OpenAI. Ona is a “boring but necessary” company that delivers a workspace for Codex that enterprises can run in their own virtual private cloud, with governance, persistence, and controls over logs, credentials, and resource access. It gives IT teams a controlled environment where they can ensure agents are properly credentialed and prevented from doing what they shouldn’t, including managing read/write protections.
AI adoption is strongest in coding, and enterprise customers have higher demands for governance and security than Codex or Claude Code can currently handle. Traditional software and cloud companies are building coding and ops infrastructure around popular models, which keeps OpenAI and Anthropic on the outside. To remain relevant, both need a stronger enterprise development story, or they risk becoming just another model that can be easily replaced
OpenAI Acquires Ona to Make AI Agents Enterprise-Ready
OpenAI has agreed to acquire Ona, a cloud development environment provider with 79 employees that was formerly known as Gitpod, in a move aimed at making agentic AI more suitable for enterprise use. The acquisition will give Codex users self-hosted sandboxes, allowing tool execution to run in infrastructure that customers control instead of OpenAI’s own services.
CIOs and CISOs have many concerns about deploying fully autonomous agents. Questions include whether an agent might delete critical files, run off on unintended tasks and generate massive token bills, or be manipulated by external actors into performing malicious actions. Ona’s technology is designed to reduce those risks by providing secure, persistent environments where agents can access tools, systems, and context over time.
What Ona Brings to OpenAI
In an OpenAI statement, the company said Ona’s platform offers trusted, customer-controlled cloud environments where work continues across devices and inside the systems where software actually lives. By integrating Ona, OpenAI plans to expand Codex beyond work tied to a single device or active session and help more organizations deploy agents securely in production.
Ona CEO Johannes Landgraf echoed that view, noting that Ona provides the building blocks agents need for enterprise work, while OpenAI brings frontier intelligence, product polish, and a scale of research and distribution that Ona could not reach alone.
Landgraf did not disclose revenue figures, but he hinted at large customers. Since the start of the year, weekly Ona agent sessions have grown 13× in production across institutions including the oldest bank in the U.S., one of Europe’s largest pharma companies, and one of Asia’s largest sovereign wealth funds.
Acquisition Price and Market Context
IDC estimates Ona’s 2025 annual revenue at roughly $7 million, with 2026 revenue likely between $10 million and $15 million. Using a standard 30× revenue multiple for acquisitions, that suggests a price of around $450–$500 million.
IDC sees this as a strategically smart move for OpenAI. While OpenAI has invested heavily in Codex, it lacked a safe, persistent area for enterprise autonomous agents. Ona provides secure environments where agents can have memory and operate under customer control—technology that is likely necessary for enterprise deployments, even if its exact quality is hard to judge.
Competitive Pressure from Anthropic
Gartner’s First Take notes that the acquisition gives Codex the scaling capability it previously lacked, but it also forces enterprises to weigh the benefits of a vendor-specific integrated stack against the flexibility of staying vendor-agnostic. Gartner also says this move appears to be OpenAI’s response to Anthropic supporting self-hosted sandboxes in Claude Managed Agents starting May 2026.
Tom Findling, CEO of Conifers.ai, sees competition with Anthropic as a major factor. He认为 this is less about removing a small competitor and more about ensuring Codex is enterprise-ready before Anthropic gains too much momentum. In the enterprise market, the battle is not just about who has the smartest coding model, but who can make AI agents safe and useful enough for large companies to deploy.
Jason Andersen of Moor Insights & Strategy agrees that OpenAI and Codex have given ground to Anthropic and Claude Code. He emphasizes that the real issue is not making Codex better at writing code, but making it work inside real enterprise environments where security, access controls, persistent cloud workspaces, audit trails, and integration with existing developer workflows are critical. Ona provides some of that missing plumbing.
Why This Matters for Enterprise Adoption
Jeremy Roberts of Info-Tech Research Group sees this as a positive step for OpenAI. Ona is a “boring but necessary” company that delivers a workspace for Codex that enterprises can run in their own virtual private cloud, with governance, persistence, and controls over logs, credentials, and resource access. It gives IT teams a controlled environment where they can ensure agents are properly credentialed and prevented from doing what they shouldn’t, including managing read/write protections.
AI adoption is strongest in coding, and enterprise customers have higher demands for governance and security than Codex or Claude Code can currently handle. Traditional software and cloud companies are building coding and ops infrastructure around popular models, which keeps OpenAI and Anthropic on the outside. To remain relevant, both need a stronger enterprise development story, or they risk becoming just another model that can be easily replaced
Archives
Categories
Archives
OpenAI Acquires Ona to Make AI Agents Enterprise-Ready
June 19, 2026Google Launches Gemini 3.5 Flash for Enterprise AI Agents
May 30, 2026Categories
Meta