Claude Code is now the most-used AI coding tool among software engineers, eight months after its May 2025 release. A Pragmatic Engineer survey of nearly 1,000 respondents, conducted January 27 to February 17, 2026, puts Claude Code at number one, ahead of GitHub Copilot and Cursor. For context: Claude Code has reached adoption levels comparable to where GitHub Copilot stood in spring 2023, its dominant era. OpenAI's Codex, which did not exist at the last survey nine months ago, already sits at 60% of Cursor's usage volume.
The adoption numbers are blunt. 95% of respondents use AI tools at least weekly. 75% apply AI to half or more of their engineering work. 56% report doing 70% or more of their work with AI. Agent usage has hit 55% overall, with staff-plus engineers leading at 63.5%. Engineers using agents are nearly twice as likely to report excitement about AI. Those not using agents are twice as likely to be skeptical. Company size shapes tool choice as much as personal preference: 75% of engineers at tiny startups use Claude Code, while 56% at 10,000-plus person companies default to GitHub Copilot, a pattern that tracks enterprise procurement over individual selection.
The full article is worth reading for the breakdown that goes beyond headlines. The chatbot race between ChatGPT, Gemini, and Claude as standalone tools is nearly a three-way tie, with no clear winner in that category. The data on seniority and tool preference is specific: directors and senior leaders favor Claude Code at twice the rate of less senior engineers, while enthusiasm for Cursor decreases as seniority increases. Full subscribers also get access to a 35-page detailed report. The market is moving fast enough that nine months of data separates the old picture from a completely different one.
[READ ORIGINAL →]