GitHub adds 4 million CPU cores and 160PB of storage as AI-generated code surges
Agent-opened pull requests now make up over 40% of activity, and GitHub has rebuilt parts of its backend to keep up.
GitHub reports that AI-driven code activity over the past year forced it to add 4 million CPU cores, 160 petabytes of storage and 50 Tb/s of networking capacity. Code commits rose sixfold and merged pull requests quadrupled year over year, with monthly commits climbing from 1.4 billion in April to 7.4 billion in September. GitHub attributes the surge to agentic AI, noting that more than 40% of pull requests were opened by an agent by last September, up from 17% in March 2026. Each commit or pull request triggers linting, tests and other automation, which strains CI/CD pipelines and pushed GitHub to try to decouple storage from compute in its backend. For developers, the takeaway is that infrastructure demand is growing faster than expected and further hardware upgrades are likely soon.
Why it matters
The growth shows that AI agents are now a large share of the work flowing through the platform millions of developers rely on, not a side experiment. For people who follow the industry, it is a concrete sign of how quickly automated code generation is changing the volume of software being written and reviewed. It also shows that the cost of that change lands on the infrastructure that hosts the code, which GitHub has had to expand and restructure in a single year.
Matters or hype?
Does this story matter, or is it hype? Vote before you see what everyone else said.