Infrastructure Expansion
GitHub, the world's largest code repository, added 4 million CPU cores, 160PB of storage, and 50 Tb/s of networking capacity over the course of the year to keep up with the rapid influx of AI-generated code.
Raw repository activity saw a massive year-over-year increase, according to the platform's technical report. Code commits involving adding or removing code grew sixfold, while merged pull requests finalizing features or bug fixes jumped fourfold.
The exponential growth pattern demonstrates rapid acceleration: monthly commits rose from 1.4 billion in April to 2.9 billion in August, reaching 7.4 billion by September. GitHub processed 17.9 billion code pushes over the past 12 months, surpassing the totals for 2024 and 2025 combined.
The Rise of Agentic AI
Agentic AI is driving the surge in code volume, forcing the company to upgrade and rearchitect its technological infrastructure. By September, over 40% of pull requests were initiated by an AI agent, up from 17% earlier in the year. Furthermore, repositories utilizing AI generate eight times the activity compared to traditional handcrafted operations.
Pipeline Strain and Hardware Demands
This code volume places significant strain on continuous integration and continuous delivery pipelines. Each commit or merged pull request triggers automated actions such as checking for remaining to-do items, code linting for visual consistency, and integration tests to ensure new code functions correctly with existing applications, services, and APIs.
Handling continuous triggers for every commit and pull request has caused hardware requirements to balloon, demanding massive computing power and input/output capacity. The intensity of the growth forced the firm to reengineer its backend architecture to decouple storage from compute, with further infrastructure upgrades anticipated as demand continues to climb.
