OpenAI Builds Pando Git Backend After Codex Agents Overload Monorepo
Original title: OpenAI Builds New Git Backend ‘Pando’ After Codex AI Agents Overwhelm Existing Infrastructure
OpenAI's Pando stores per-branch packfiles as immutable blobs to cut median push latency by 5x for AI agent workloads.
OpenAI deployed Pando after Codex AI agents increased monorepo branch pushes and fetches. It uses metadata stores for ref versions and local disk cache for packfiles. The system served over 57 TB of data in one hour with 99% cache hits. OpenAI built Pando in 10 weeks using agents and contributed upstream Git fixes for concurrent downloads.
- Median push latency fell by nearly 80% after cutover.
- Pando advertises only the trunk branch by default.
- Small median packfiles under 20 KiB enable logarithmic compaction.
From the original article
3 min read at gbhackers.com
OpenAI has built a new Git backend called Pando after Codex-driven development overwhelmed the systems behind its internal monorepo. OpenAI’s engineering post, published November 2, 2026, says Pando now powers all monorepo development. It cut median push latency by nearly 5x and delivers nearly a petabyte of Git data per day.
By Divya · gbhackers.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.