OpenAI to acquire Pierre Computer and fold its Code Storage system into Codex
OpenAI has announced plans to acquire Pierre Computer Company, the developer-tools startup whose Git-based Code Storage service and open-source Trees and Diffs libraries already power parts of its Codex coding app.
OpenAI has announced plans to acquire Pierre Computer Company, a small developer-tools startup whose technology already runs inside its Codex coding app, in a deal that sharpens the contest over how software is written when artificial intelligence agents do much of the typing.
Under the planned acquisition, Pierre’s commercial product, Code Storage, will be folded into Codex, OpenAI’s AI coding agent. Code Storage is a Git-based hosting service built for machines rather than only for human developers: applications can create repositories, push code and track its history through an application programming interface, without the rate limits and sign-in steps of services designed for people. Pierre launched it in October 2025, describing it as a low-latency Git infrastructure layer that other companies can embed in their own apps.
The financial terms of the deal were not disclosed. OpenAI said the acquisition is subject to customary closing conditions, including regulatory approvals, and that until it closes the two companies will remain separate and independent. Pierre’s open-source projects will remain open source, OpenAI said, and existing Code Storage customers will be able to keep using the service while its features are brought into Codex.
What makes the deal notable is how much of Pierre is already inside Codex. OpenAI said anyone who has browsed a project’s files or reviewed a change in the Codex app has already used Pierre’s Trees library, which powers file navigation, and its Diffs library, which renders changes. A third open-source project, DiffsHub, completes a family of tools concerned with the most human part of AI-assisted coding: reading, reviewing and understanding what an agent has just written.
That focus explains why observers have framed the acquisition as a move against Microsoft-owned GitHub, long the central place where developers store and collaborate on code. As coding agents generate more code, more quickly, the bottleneck is no longer writing it but storing, versioning, reviewing and rolling it back reliably. A coding assistant that also owns the storage layer can offer that whole loop in one place, rather than handing code off to a separate platform at the end of each task.
Pierre says its service is already used by AI companies including Lovable, Bolt, Amp and Poke, which use it to store generated code as well as agent memory files, design files and other documents that change over time. Its published specifications describe a distributed system built on Git’s commit protocol, with support for very large repositories and Git Large File Storage, aimed at workloads where dozens of writes can arrive at once from parallel agents.
The announcement landed just two days after OpenAI described Pando, its own in-house backend for serving Git repositories, which it said now powers development in its internal monorepo and cut median push latency by nearly 80 per cent. The two systems address related but distinct problems: Pando serves OpenAI’s own engineers and agents at scale, while Code Storage is a product other companies embed. OpenAI has not said it will merge them into a single public service.
The Pierre deal also arrived on the same day as a parallel move in the AI coding race: Meta’s acquisition of Conductor, the startup whose app runs teams of coding agents in parallel. Together, the two deals show the largest AI companies buying not just models but the surrounding workflow — the file trees, diff viewers, review layers and storage systems through which developers and agents actually work together.
For developers, the immediate practical questions are about continuity and choice. OpenAI has promised that Code Storage will keep running for existing customers and that its open-source libraries will stay open, which matters to teams that built on Trees and Diffs independently of Codex. The longer question is whether coding infrastructure consolidates around a few AI platforms, or whether open protocols and syncing with GitHub, GitLab and Bitbucket — which Pierre already supports — keep the ecosystem open.
No closing date has been given, and regulatory review could still shape the outcome. But the direction of travel is clear: as AI agents take on larger programming jobs, the companies building them are racing to own the unglamorous machinery underneath — storage, version history and review — on which the whole agent-coding boom depends.
More in Technology & AI
Photo: Meta headquarters in Menlo Park, California (file image)
Technology & AIMeta acquires AI coding startup Conductor and winds down its cloud service
Meta has acquired Conductor, the startup whose app runs teams of AI coding agents in parallel, with the team joining Meta Superintelligence Labs and Conductor Cloud shutting down on 20 November.
Photo: A silicon wafer used in semiconductor chip manufacturing (file image)
Technology & AIEU to invest €525 million in chip startups in bid to close the gap with the US and China
The European Union will put €525 million into chip startups, including €440 million in grants and €85 million for a new European Chips Design Platform, as Brussels tries to strengthen a sector where its design firms hold less than 1 per cent of global revenue.
Photo: Tencent headquarters (file image)
Technology & AITencent releases Hy4 AI model with cybersecurity training limited to defensive uses
Chinese technology company Tencent has released Hy4, a new open-source artificial intelligence model whose cybersecurity training was deliberately limited to defensive work, in an effort to reduce the risk that it could help carry out cyberattacks.