OpenAI has built Pando, a new backend for Git repositories designed to handle the surge in development activity driven by its Codex coding agent. The company says Pando now powers all development in its monorepo, and it was deployed without interrupting engineering work or changing how staff use Git. The move followed a period beginning around the end of last year when slow fetches and pushes, failed merges, and reliability at times below 90 percent strained the existing systems. Engineers were running several changes through CI at once, and each push could trigger several jobs fetching the same commit. Pando is built around two goals: horizontally scalable writes for concurrent pushes to many branches, and high read throughput so clients can download data fast enough to saturate their network bandwidth. It coordinates updates to each branch independently, so unrelated branches can be updated in parallel. It also reuses packfiles, Git's compressed bundles of objects, across fetches, so the server can send the same bytes to many clients with little extra computation. The trade-off is that a reader looking at several branches may see them at different points in time. OpenAI reports that in production, Pando cut median push latency by nearly five times and now serves nearly a petabyte of Git data per day. Because it keeps the Git protocol and client, engineers and agents can adopt it without changing their workflows. OpenAI says it will also contribute its client-side improvements upstream to the broader Git project. For Australian businesses and professionals, this changes little in practical terms. Pando is an internal system described in the article, and nothing in the post indicates it is available as a product. Teams using standard Git hosting will see no change in cost, risk or regulatory obligations, and no Privacy Act action is triggered by this announcement. The most relevant takeaway is that agent-driven development is putting real load on source control, which may matter for organisations planning large monorepos or heavy CI use. For most teams, there is nothing new to adopt yet.