GitHub CEO Thomas Dohmke: 150M developers, 1B repos, and Copilot driving the next platform shift

Summary

GitHub has crossed 150 million developers and 1 billion repositories on its platform, while GitHub Copilot has reached 20 million users and the business surpassed $2 billion in annual recurring revenue roughly a year ago. Copilot now reaches 90% of the Fortune 100, and Thomas Dohmke, who has led GitHub for four years following Microsoft's 2018 acquisition, frames these as compounding network effects rather than a single product win.

The growth runway for Copilot remains substantial. Microsoft's Visual Studio family, comprising VS Code and VS Enterprise, has 50 million active users, meaning Copilot has penetrated less than half that addressable base even among developers already inside the Microsoft ecosystem. Dohmke also notes that fewer than 50% of GitHub's 150 million developers use AI tools on a daily basis, citing students, legacy-stack maintainers, and late adopters as the unconverted segment.

Where enterprise value is actually being captured is less exotic than the agentic narrative suggests. Dohmke identifies in-IDE assistance, including code completion and contextual chat, as still the primary value driver for Fortune 100 customers. Agentic workflows, where the tool autonomously installs packages, reads error logs, and writes test cases, are growing but sit at the far end of a spectrum where developers remain in control by design. His view is that most developers want AI to amplify their work, not replace their agency.

Model agnosticism is now a strategic pillar. Since last year, Copilot has offered multimodel choice spanning OpenAI, Anthropic, and Google, plus a bring-your-own-model path via OpenRouter and Llama-compatible APIs. Dohmke frames model lock-in as a non-starter for any developer tool competing for market share, drawing a direct parallel to GitHub's language and library neutrality as foundational to its scale. The practical constraint is integration capacity: new models are shipping daily, each requiring GPU provisioning, responsible AI testing, and red-teaming before deployment.

The longer-term platform thesis Dohmke articulates positions GitHub as the collaboration layer not just between human developers, but between humans and agents, and eventually between agents themselves. That framing aligns GitHub's trajectory with Microsoft's broader Azure AI Foundry investments and suggests the repository and identity infrastructure GitHub controls becomes more strategically valuable as agentic software development scales.