OpenAI to acquire Ona

Overview

OpenAI plans to acquire Ona, the German cloud platform formerly known as Gitpod, to give its Codex coding agent secure environments where work keeps running for hours or days after a developer logs off. Ona has supported about 2 million developers in reproducible cloud sandboxes. Financial terms were not disclosed, and the deal is subject to regulatory approval.

Key Takeaways

OpenAI will fold Ona into its Codex coding team to run AI agents inside secure, persistent cloud sandboxes rather than on a developer's laptop.
Ona moves agent work into cloud environments that stay online after a workstation shuts down, so long jobs finish without interruption.
Agents run inside the customer's own cloud under Ona's execution model, while OpenAI supplies the models and orchestration.
Codex now reaches more than 5 million weekly users, up about 400% from earlier in 2026, driving the need for longer-running capacity.
Financial terms were not made public, and the purchase still needs customary regulatory approvals before it closes.

Stats & Key Facts

#About 2 million developers have worked in Ona's secure, reproducible cloud environments.
#More than 5 million people now use Codex each week to research, build, and automate work.
#Codex weekly usage is up about 400% from earlier in 2026, rising from roughly 3 million users in April.
#Ona is OpenAI's sixth acquisition in 2026 by June, close to its full-year total of eight in 2025.
#Ona, founded as Gitpod in Kiel, Germany, has operated since around 2020 before its 2025 rebrand.

Why OpenAI Is Buying Ona to Extend Codex

The purchase targets a specific limit in how coding agents work today.

As coding agents take on bigger jobs, the most valuable work now unfolds over hours or days instead of minutes. OpenAI argues that work should not stop the moment a developer closes a laptop. Ona supplies the cloud layer that keeps those tasks alive.

Rather than spend years building its own cloud orchestration, OpenAI is buying infrastructure that already runs at scale. The bet is that purpose-built agent environments will set production coding tools apart from prototyping toys.

How Ona's Persistent Cloud Sandboxes Keep Work Running

Ona moves agent workloads off local machines and into the cloud.

›Each sandbox is pre-configured with the tools, systems, and context an agent needs to finish a task.
›Sandboxes stay online after a developer's workstation shuts down, so long-running jobs continue.
›Environments are reproducible, meaning a task starts the same way every time.
›Sandboxes are removed automatically once the assigned work is complete.

Customer-Controlled Execution and Security Guardrails

Security is the central selling point for business buyers.

Ona's customer-controlled execution model lets agents operate inside an organization's own cloud, while OpenAI provides the intelligence and orchestration. That keeps sensitive code and data within the company's perimeter.

Built-in guardrails block agents from reading file paths that hold encryption keys and credentials. Outbound connections to suspected malicious servers are blocked. Ona also identifies blocked applications by a cryptographic signature, so a flagged program is caught regardless of its file name or location.

Codex Growth and the Push Toward Bigger Jobs

The deal lands as Codex usage climbs sharply.

›More than 5 million people now use Codex each week to research, analyze, build, and automate work.
›Weekly usage is up about 400% from earlier in 2026, rising from roughly 3 million users in April.
›OpenAI frames Ona as the way to handle longer, more ambitious delegated tasks at this scale.

Ona's History as Gitpod and Its Enterprise Roots

The company brings an established developer base to OpenAI.

Ona started as Gitpod, founded around 2020 in Kiel, Germany, and rebranded in 2025 as it rebuilt itself around AI agents. Its cloud workspaces have supported about 2 million developers.

Reports describe an enterprise customer mix spanning regulated industries, the kind of buyers that demand strict security and audit controls. OpenAI and Ona already share multiple customers, which should ease the path to a combined product.

Competitive Pressure From Anthropic in Enterprise Coding

The acquisition is also a competitive response.

Anthropic has gained ground with enterprise coding workflows, and its Claude Code tool is widely seen as a leader for extended coding tasks. Buying Ona gives OpenAI a secure cloud foundation to challenge that position.

Ona is OpenAI's sixth acquisition in 2026 by June, close to its full-year total of eight in 2025. An earlier 2026 deal brought the Python tools uv and Ruff into the Codex ecosystem, showing a pattern of buying developer infrastructure rather than building it.

Deal Terms, Timing, and What Stays Independent

Key conditions remain open.

›Financial terms were not made public.
›The acquisition is subject to customary closing conditions, including required regulatory approvals.
›OpenAI and Ona will operate as separate, independent companies until the deal closes.
›The Ona team will join OpenAI's Codex group to build the integration.

Frequently Asked Questions

What does Ona do?

Ona runs AI coding agents inside secure cloud sandboxes that stay online after a developer logs off, so long tasks finish without interruption. Each sandbox comes pre-configured with the tools and context an agent needs.

Was Ona formerly called Gitpod?

Yes. The company began as Gitpod, founded around 2020 in Kiel, Germany, and rebranded to Ona in 2025 as it shifted its focus to AI agents.

How much did OpenAI pay for Ona?

OpenAI did not disclose the financial terms. The deal still needs customary regulatory approvals before it closes, and the two firms stay independent until then.

How does Ona keep enterprise data secure?

Agents run inside the customer's own cloud, and guardrails block them from reading files holding credentials and encryption keys. Outbound connections to suspected malicious servers are also blocked.

Why does this matter for Codex users?

It lets work delegated to Codex keep running in a controlled cloud environment for hours or days, rather than staying tied to the machine where it started. That supports bigger, longer coding jobs.

The Ona deal gives OpenAI a secure cloud home for long-running Codex agents and a stronger hand against rivals in enterprise coding. The purchase still needs regulatory clearance before it closes.