OpenAI just announced it’s acquiring Ona, and if you only read the headline, you might think this is just another acqui-hire. It’s not. When OpenAI acquires Ona, it’s buying the one piece Codex was missing: infrastructure that lets AI agents keep working after you close your laptop.
What Actually Happened
OpenAI has entered an agreement to acquire Ona, a company that builds secure cloud execution environments for AI agents. The deal, announced June 11, 2026, will fold Ona’s team and technology into OpenAI’s Codex division.
Financial terms weren’t disclosed. The acquisition still needs regulatory approval, so nothing’s final yet. Until it closes, both companies keep operating independently.
If the name Ona doesn’t ring a bell, you might know its previous life: Gitpod. The company rebranded to Ona, but the sign-in page still routes to app.gitpod.io. Ona spent years helping developers move software work from local machines into the cloud, and that experience is exactly what OpenAI wants.
What Ona Actually Does
Ona builds what it calls “the platform for background agents.” Here’s what that means in practice:
Agents that don’t die when you close your laptop. Right now, every AI coding tool is tied to your active session. Close the tab, shut the lid, the work stops. Ona’s tech gives agents persistent cloud environments where they keep running.
Your infra, your rules. Ona deploys inside your VPC (Virtual Private Cloud). You control network access, credentials, permissions, and audit trails. The agent works within your security boundaries, not some shared sandbox.
Enterprise-grade governance. Kernel-level security policy enforcement, scoped credentials, activity logging, and review workflows. This is the stuff enterprises need before they’ll trust agents with real work.
Ona’s been doing this at scale. They’ve served 2 million developers and count BNY (the oldest bank in the US), European pharma companies, Asian sovereign wealth funds, and firms like GSR, Pearson, EquipmentShare, and Hargreaves Lansdown as customers.
The growth numbers are wild. Weekly Ona agent sessions jumped 13x since the start of 2026.
Why OpenAI Wanted This
Codex is having a moment. Over 5 million people use it weekly, up 400% from earlier this year. OpenAI’s been building out Codex fast, recently publishing a use-case hub with 60+ workflows. But here’s the thing: Codex started as a developer tool, and it’s expanding fast. Non-developers (analysts, marketers, investors, bankers) now make up about 20% of Codex users and are growing three times faster than developers.
That growth creates a problem. As Codex gets used for more complex work, the jobs take longer. We’re talking hours or days, not minutes. And you can’t ask someone to keep a browser tab open for 16 hours while an agent modernizes a codebase or runs security scans.
Ona solves this. Its technology gives Codex a persistent place to work: secure cloud environments where agents can access tools, systems, and context over time, even when nobody’s watching.
“Agents need more than intelligence; they need a trusted workspace,” said Johannes Landgraf, Ona’s co-founder and CEO. “We built Ona to give agents cloud environments with the context, control, and collaboration enterprises require.”
Thibault Sottiaux, OpenAI’s Core Products Lead, put it this way: “Enterprises want powerful agents that can do real work while meeting the security and control requirements of their environments. Ona will help us make Codex easier to deploy securely across production workflows.”
What the Ona Acquisition Changes If You Use Codex
If you’re using Codex for anything beyond quick one-off tasks, this acquisition matters. Here’s why:
Work can run in the background. Hand off a complex task, close your laptop, check back later. The agent keeps going in the cloud. This is the “fire and forget” mode Ona customers already use.
It’s not just for coding anymore. OpenAI has been pushing Codex beyond software development. Recent plugins target public equity, banking, and sales, with legal, marketing, and corporate finance coming. Persistent execution makes all of those viable for longer-running work.
Enterprise IT can actually say yes. The security and governance features (VPC deployment, audit trails, scoped credentials) are what enterprise security teams need to approve agent deployment at scale. Without them, agents stay in the “interesting experiment” category.
This is the difference between a tool that demos well and one you can trust with production workloads.
The Bigger Picture
This acquisition fits into a much larger pattern. OpenAI is aggressively pushing into enterprise. The company recently filed a confidential S-1 for its IPO, and CFO Sarah Friar said in January she expects enterprise to grow from 40% to 50% of OpenAI’s revenue by end of year.
The Visa partnership for AI agent payments (announced the same week) is another signal. OpenAI wants agents that don’t just write code but handle real business workflows, including financial transactions.
Codex expanding beyond developers is the third piece. If 20% of users are already non-developers and that segment is growing 3x faster, the “AI workforce” framing Ona uses isn’t aspirational. It’s where things are heading.
What to Watch
Regulatory approval. The deal needs it. No timeline given. Could be fast, could drag.
Integration details. How does Ona’s platform merge with Codex? Does the Gitpod brand disappear? What happens to existing Ona customers?
Competitor response. Cursor, Windsurf, and other AI coding tools don’t have their own cloud execution layer. Will they build, buy, or partner?
Pricing. Will persistent agent execution be included in existing Codex plans, or is this a premium tier? Nobody’s said yet.
The Bottom Line
OpenAI didn’t just buy a company. It bought the infrastructure layer that makes AI agents trustworthy for real work. Ona’s tech turns Codex from “cool tool that works when you’re watching” into “agent that handles overnight jobs on your infra with your security rules.”
That’s a meaningful shift. And if you’re building anything with AI agents, it’s worth paying attention to.
Source: OpenAI announcement (June 11, 2026), Ona CEO blog post (June 11, 2026), PYMNTS coverage, Yahoo Finance reporting. All growth figures and user counts are company-sourced and not independently verified.
Editorial note: This article is based on publicly available announcements and reporting. Tony Reviews Things has no commercial relationship with OpenAI or Ona.
