--- title: Scaling Git at OpenAI | daily.dev description: OpenAI built Pando, a new Git-compatible backend to handle the scale of concurrent development Codex enables. Pando cut median push latency by nearly 80% and... canonical: https://daily.dev/posts/scaling-git-at-openai-mksou1src twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Scaling Git at OpenAI | daily.dev og:description: OpenAI built Pando, a new Git-compatible backend to handle the scale of concurrent development Codex enables. Pando cut median push latency by nearly 80% and... og:url: https://daily.dev/posts/scaling-git-at-openai-mksou1src og:image: https://api.daily.dev/og/posts/mKSOu1src.png og:image:alt: Scaling Git at OpenAI og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Scaling Git at OpenAI **[OpenAI](https://daily.dev/sources/openai)** · 25 min read · 0 upvotes · 0 comments ## Summary OpenAI built Pando, a new Git-compatible backend to handle the scale of concurrent development Codex enables. Pando cut median push latency by nearly 80% and now delivers close to a petabyte of Git data per day. It uses causal consistency instead of repository-wide consistency to let writes to unrelated branches proceed independently, and reuses cached packfiles instead of generating one per client to scale reads. Pando has served over 57 terabytes of Git packs in a single hour, with more than 99% from local disk cache. It was built in Rust, went from kickoff to production in 10 weeks, and began serving production traffic in April. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### How much did OpenAI's Pando backend reduce Git push latency compared to their old system? Pando reduced the median push latency by nearly 80% compared to OpenAI's previous Git backend. It also cut median push latency by nearly 5x in production overall, while improving merge reliability. Pando now delivers nearly a petabyte of Git data per day, and has served more than 57 terabytes of Git packs in a single hour, with over 99% served from local disk cache. _Teams hitting Git scaling limits can track emerging source-control architectures like this one on daily.dev._ ### What consistency model does Pando use instead of strict repository-wide consistency for Git? Pando uses causal consistency rather than a single global ordering of all ref updates. Updates to each branch are coordinated independently, so pushes to unrelated branches proceed in parallel, while clients pushing to the same ref still get serializable ordering. A client reading several branches at once may see them at different points in time rather than one consistent snapshot. _Engineers evaluating consistency tradeoffs in distributed systems can follow write-ups like this on daily.dev._ ### What programming language did OpenAI use to build their new Git backend Pando? OpenAI implemented Pando in Rust, choosing it as a greenfield project without legacy dependencies. The core design was done by a single engineer, implemented by teams of agents using Codex, and deployed by a handful of engineers, taking the project from kickoff to production in about 10 weeks. _Developers weighing Rust for new infrastructure projects can track real-world build reports on daily.dev._ ## Similar posts on daily.dev - [Building Git infrastructure for agent-scale development](https://daily.dev/posts/building-git-infrastructure-for-agent-scale-development-xo12uijhx) · GitHub Blog · 7 upvotes · 0 comments - [Pushing 4.7M Commits an Hour](https://daily.dev/posts/pushing-4-7m-commits-an-hour-kmqolalru) · Matt Rickard · 0 upvotes · 0 comments - [Redesigning source control at Anthropic](https://daily.dev/posts/redesigning-source-control-at-anthropic-h0zfzbh1m) · Reid Burke · 0 upvotes · 1 comments --- Tags: [#openai](https://daily.dev/tags/openai), [#rust](https://daily.dev/tags/rust), [#git](https://daily.dev/tags/git), [#distributed-systems](https://daily.dev/tags/distributed-systems), [#openai-codex](https://daily.dev/tags/openai-codex) [View this post on daily.dev](https://daily.dev/posts/scaling-git-at-openai-mksou1src) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Scaling Git at OpenAI","url":"https://daily.dev/posts/scaling-git-at-openai-mksou1src","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/scaling-git-at-openai-mksou1src"},"datePublished":"2026-11-02T23:52:03.921Z","dateModified":"2026-11-03T02:34:06.346Z","description":"OpenAI built Pando, a new Git-compatible backend to handle the scale of concurrent development Codex enables. Pando cut median push latency by nearly 80% and...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/bbb3ffce38b3fc0c63735f9434465b8e?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/bbb3ffce38b3fc0c63735f9434465b8e?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"OpenAI","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"OpenAI","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/984b3632316f4afaa3d5213d5c64fe65","url":"https://daily.dev/sources/openai"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/scaling-git-at-openai-mksou1src","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"openai,rust,git,distributed-systems,openai-codex","timeRequired":"PT25M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"OpenAI","item":"https://daily.dev/sources/openai"},{"@type":"ListItem","position":3,"name":"Scaling Git at OpenAI"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/scaling-git-at-openai-mksou1src#faq","mainEntity":[{"@type":"Question","name":"How much did OpenAI's Pando backend reduce Git push latency compared to their old system?","acceptedAnswer":{"@type":"Answer","text":"Pando reduced the median push latency by nearly 80% compared to OpenAI's previous Git backend. It also cut median push latency by nearly 5x in production overall, while improving merge reliability. Pando now delivers nearly a petabyte of Git data per day, and has served more than 57 terabytes of Git packs in a single hour, with over 99% served from local disk cache. Teams hitting Git scaling limits can track emerging source-control architectures like this one on daily.dev."}},{"@type":"Question","name":"What consistency model does Pando use instead of strict repository-wide consistency for Git?","acceptedAnswer":{"@type":"Answer","text":"Pando uses causal consistency rather than a single global ordering of all ref updates. Updates to each branch are coordinated independently, so pushes to unrelated branches proceed in parallel, while clients pushing to the same ref still get serializable ordering. A client reading several branches at once may see them at different points in time rather than one consistent snapshot. Engineers evaluating consistency tradeoffs in distributed systems can follow write-ups like this on daily.dev."}},{"@type":"Question","name":"What programming language did OpenAI use to build their new Git backend Pando?","acceptedAnswer":{"@type":"Answer","text":"OpenAI implemented Pando in Rust, choosing it as a greenfield project without legacy dependencies. The core design was done by a single engineer, implemented by teams of agents using Codex, and deployed by a handful of engineers, taking the project from kickoff to production in about 10 weeks. Developers weighing Rust for new infrastructure projects can track real-world build reports on daily.dev."}}]} ```