OpenAI to buy Pierre Computer, folding Git-based code storage into its Codex agent

OpenAI has announced plans to acquire Pierre Computer Company, a developer-tooling startup whose Git-based code storage service and open-source libraries already underpin parts of the company's Codex coding agent. The terms of the deal were not disclosed, and the transaction is subject to approval by regulators. The move marks an extension of the AI coding race beyond model capability into the repositories where agent-written code is stored, reviewed and merged — infrastructure that has until now sat largely with existing code-hosting providers.

According to OpenAI's announcement, the Pierre team will work with the company to "advance the Git workflows developers already use" and to tune them to the way individual teams collaborate with coding agents. OpenAI said it wants to set a standard for modern software development "from the moment an agent creates its first branch to the moment code is deployed to the world." Two of Pierre's open-source libraries, Trees and Diffs, are already used in the Codex app for file exploration and change review; SiliconANGLE reported that Trees helps implement folder navigation sidebars with search, drag-and-drop controls and flags on recently changed files, while Diffs identifies the specific lines altered in an updated file. Viewed from the industry, observers cited by SiliconANGLE characterised the purchase as a bid to take on Microsoft's GitHub in the era of AI agents, and experts cited by AI Times described it as an attempt to challenge GitHub's long-standing role at the centre of code storage and developer collaboration.

Under the plans described, OpenAI will continue to offer Trees and Diffs under an open-source licence once the acquisition closes, and intends to integrate the features of Pierre's paid code-hosting service into Codex. Pierre's open-source projects are to remain open source, and customers of its code storage service will be able to keep using existing services while that integration proceeds. Pierre says its service performs significantly better than competing technologies and helped one customer, the AI startup Lovable, cut code retrieval latency by about 70%.

The acquisition is presented against a background of strain on conventional code management as AI agents take a larger part in development. According to AI Times, multiple agents writing code simultaneously multiply branches and revision histories, creating bottlenecks in storage and merging. OpenAI disclosed in a blog post that it had built its own source management system, Pando, designed to handle concurrent work across many branches without changing how developers use Git; the company said median code push latency fell by about 80% after its introduction, with roughly one petabyte of Git data transferred daily. The financial value of the Pierre deal remains undisclosed, and no completion date has been given; the transaction now awaits regulatory clearance.