GitHub Adds 4 Million CPU Cores as AI Code Pushes Commits to 7.4 Billion a Month

GitHub says a surge in AI-generated code forced it to add 4 million CPU cores, 160 petabytes of storage and 50 Tb/s of networking capacity. The company reported that monthly code commits rose to 7.4 billion in September, a sixfold increase over the prior year. The growth also pushed GitHub to reengineer parts of its backend.

Repository activity surged

According to GitHub's technical report, code commits grew sixfold year over year, while merged pull requests jumped fourfold. In April, there were 1.4 billion commits per month; that figure rose to 2.9 billion in August and reached 7.4 billion in September. GitHub also saw 17.9 billion code pushes over the past 12 months, which it said is more than all of 2024 and 2025 put together.

Agentic AI drove the increase

GitHub said agentic AI is behind the rise in code volume and the need to upgrade and rethink its infrastructure. By last September, more than 40 percent of pull requests were started by an agent, up from 17 percent in March 2026. Repositories that begin with AI generally see eight times the activity of handcrafted operations, the company said.

CI/CD pipelines face new strain

All that activity puts disproportionate strain on CI/CD pipelines. Each commit or merged pull request can trigger checks for leftover to-do items, linting for consistent formatting, and code and integration tests. GitHub said the rise was intense enough to require reengineering its backend to decouple storage from compute as much as possible, and it is likely to need further upgrades soon.

Sources: Tom's Hardware