OPENAI ACQUIRES ONA — KEY FACTS
● Target: Ona — formerly Gitpod, a German cloud development environment company
● Purpose: Enable Codex to run as persistent cloud-based agents rather than session-limited local processes
● Codex users: Now exceeds 5 million weekly active users
● What Gitpod/Ona does: Cloud development environments — spins up full dev environments in the cloud, no local setup required
● What this unlocks for Codex: Long-running coding tasks that persist across sessions, don't require local machine to stay active, can run in parallel at scale
● Competitive context: Claude Code already runs persistent agentic sessions; OpenAI is closing the architecture gap
● The signal: OpenAI is treating Codex as a first-class product investment — an acquisition, not just a feature update
Why Persistent Cloud Agents Change the Coding Agent Market
OpenAI acquired German startup Ona, formerly known as Gitpod, to enhance its Codex coding tool by enabling persistent cloud-based agent operation. The acquisition aims to support Codex's growing user base, which now exceeds five million weekly users, amid fierce competition in the AI coding market. The architecture problem Ona solves is fundamental to how agentic coding agents work today. Current coding agents run as processes on a user's local machine or in a temporary cloud session. When the session ends — when the user closes their terminal, when a timeout occurs, when a machine restarts — the agent stops. Long-horizon coding tasks that require hours of continuous execution must be manually restarted or structured to fit within session limits.
Gitpod's original product was cloud development environments: instead of setting up a local dev environment, developers could open a URL and get a fully configured, running development environment in a browser within seconds. Ona (the company's current name) extended this to support persistent, cloud-native development workflows. Applied to Codex, this means: assign a coding task, close your laptop, and the Codex agent continues running in a cloud environment until the task is complete or encounters a blocker that requires human input. That is the core use case that differentiates a coding agent from a coding assistant.
5 Million Weekly Users — The Codex Growth Story
Codex's 5 million weekly active users makes it one of the fastest-growing developer tools in history. For context: GitHub Copilot reached 1 million users in its first year (2022) and now has approximately 1.8 million paid subscribers. Cursor reached approximately 4 million active users by mid-2026 before the SpaceX acquisition. Codex at 5 million weekly users — in a product that only launched publicly in May 2026 — represents a growth rate that validates OpenAI's decision to treat it as a standalone product rather than a ChatGPT feature.
The acquisition investment — terms not disclosed — is directionally consistent with a product OpenAI sees as a $1B+ revenue opportunity. If Codex converts even 5% of its weekly active users to paid tiers at $20-30/month, that is $60-90 million in monthly recurring revenue — a meaningful contribution to OpenAI's path to profitability ahead of its Q4 2026 IPO.
What This Means for the Coding Agent Competitive Landscape
For Codex users: Persistent cloud agents mean you can assign a large refactor, a test suite generation, or a documentation pass and leave it running overnight without keeping your machine active. This is the feature that professional developers have been waiting for — agentic coding that matches how actual large-scale development work happens.
For Claude Code: Anthropic already supports persistent agentic sessions through Claude Code's architecture, but the Ona acquisition signals that OpenAI is investing heavily to close the architecture gap. The competitive differentiation on persistence narrows. Claude Code's differentiation shifts increasingly to SWE-bench Pro accuracy (80.4% vs Codex's unpublished score) and the Anthropic safety profile.
For Cursor post-SpaceX acquisition: Cursor's core value proposition is a full IDE with MCP integrations. Persistent cloud Codex agents are a different product category — they run headlessly on cloud infrastructure, not inside an IDE. The overlap is partial, not complete. But as persistent Codex agents mature, the question of whether developers need a local IDE at all becomes more pointed.
For the coding agent market overall: The Ona acquisition is the third major consolidation move in the AI coding market in July 2026 alone — after SpaceX's $60B Cursor acquisition and Anthropic's Ode forward-deployed engineering launch. The standalone IDE and local coding agent market is consolidating rapidly. The next 12 months will determine whether persistent cloud agents (Codex + Ona) or local IDE agents (Claude Code, Cursor + SpaceX) win enterprise developer adoption.
Sources: aibusiness.com · dentro.de/ai July 2026 roundup · Related: Kimi Code vs Codex vs Claude Code comparison → · Grok Build vs Claude Code → · Cursor Sand agent — what SpaceX may kill →