OpenAI Buys Cloud Startup Ona So Codex Can Run Tasks While Your Laptop Is Closed

The fundamental limitation of today's AI coding assistants is that they stop working when you do. Close your laptop, and your agent's context evaporates. OpenAI's acquisition of Ona — announced Thursday — is designed to fix exactly that. The German cloud startup builds secure, persistent environments where AI agents can keep working on multi‑step tasks across hours or days, even when the developer who kicked off the work has gone to sleep.

OpenAI announced the deal alongside the news that Codex now has more than 5 million weekly active users, up from 3 million in April — a roughly 400% increase from earlier this year. Financial terms were not disclosed. The acquisition is subject to regulatory approvals, and Ona's team will join OpenAI's Codex division once it closes, according to Quartz.

Ona — founded in 2019 as Gitpod before rebranding — provides pre‑configured cloud environments with access controls, audit trails, and the tools AI agents need to operate independently. Its enterprise clients include a major U.S. bank, European pharmaceutical companies, and Asian sovereign wealth funds, according to MLQ News. Productive use of Ona's agents among enterprise clients has increased 13‑fold in 2026 alone.

"Agents need more than intelligence; they need a trusted workspace," Ona CEO Johannes Landgraf said in a statement. "We built Ona to give agents cloud environments with the context, control and collaboration enterprises require."

Codex has become one of OpenAI's fastest‑growing products. The 5 million weekly active user figure represents a more than 400% increase from early 2026. But the user base is shifting: knowledge workers now make up roughly 20% of Codex users and are growing at triple the rate of software developers, according to MLQ News citing Yahoo Finance.

Thibault Sottiaux, OpenAI's Head of Core Products, said enterprises want "powerful agents that can do real work while meeting the security and control requirements of their environments." Ona will help "make Codex easier to deploy securely across production workflows," he told Stocktwits/Yahoo Finance.

Most AI coding tools today operate in short, session‑based interactions. You prompt, the agent responds, the context window fills up, and eventually you start fresh. That model breaks down for the kind of work that actually ships software: vulnerability scanning across entire codebases, multi‑day application modernization projects, automated testing pipelines, and long‑running data migrations.

Ona's persistent environments allow agents to maintain state across extended workflows, picking up complex tasks across hours or days without losing context. Once integrated, OpenAI said the combined teams will focus on helping engineering workflows tackle long‑term software lifecycle tasks like automated vulnerability scanning, application modernization, and testing.

The acquisition comes amid an intensifying AI coding war. Anthropic's Claude Code has been growing fast, and both companies have filed confidential IPO prospectuses with the SEC. OpenAI is also expanding Codex beyond pure coding — the company recently partnered with Visa to let AI agents handle financial transactions, signaling ambitions to anchor Codex deeper into business workflows.

Ona's infrastructure addresses a concrete bottleneck: as AI models get better at coding, the limiting factor isn't model intelligence — it's the operational infrastructure that lets agents do sustained work without a human keeping the laptop open. OpenAI's bet is that owning that infrastructure layer, rather than renting it, creates a moat that pure model providers can't easily cross.

Sottiaux's comment was reported by Stocktwits/Yahoo Finance in coverage of the acquisition. For individual developers, the immediate impact is narrow — the acquisition hasn't closed yet, and Ona's features aren't integrated into Codex. But the direction is clear: OpenAI is building toward Codex as a persistent agent that can be handed a multi‑hour task and trusted to execute it in the cloud, reporting back when it's done. That's a fundamentally different product than a session‑based coding assistant.

For teams and enterprises, Ona's customer‑controlled execution model is the bigger story. Organizations will be able to run Codex agents inside their own cloud environments while OpenAI provides the intelligence and orchestration. That means control over where agents run, what they can access, how credentials are scoped, and how activity is logged — the kind of governance that regulated industries require before they'll let AI agents anywhere near production systems.