🟢 READ | ⏱ 12 min | 📡 8/10 | 🎯 Software engineers, engineering leaders, CTOs

TL;DR

Pragmatic Engineer surveyed 900 engineers (Jan-Feb 2026) on AI tool usage. Claude Code has exploded to #1 in just 8 months, 95% of engineers use AI weekly, and 55% now regularly use AI agents. Company size drives tool choice: tiny startups overwhelmingly pick Claude Code (75%), while large enterprises default to GitHub Copilot (56%+).

Signal

• Claude Code adoption rocketed from zero to market leader in 8 months; 46% of engineers "love" it—2x the love for Cursor and 5x for Copilot
• AI is now mainstream: 75% of respondents do ≥50% of work with AI; 56% do ≥70%—fundamental shift from 2023's "experimental" phase
• Agent usage correlates strongly with positivity: users of AI agents are 1.7× more enthusiastic (61% vs 36%) and 2× less skeptical than non-users

What They're NOT Telling You

Survey bias: respondents are Pragmatic Engineer subscribers (already AI-forward), not representative of all engineers globally—adoption likely lower in traditional enterprises. Missing: cost analysis (pricing wars between Claude Code, Copilot, Cursor) and latency/reliability comparisons that would matter to production teams. No deep breakdown of what types of work agents handle best vs where they fail—only broad categories like "bug fixing" and "code review."

Trust Check

Factuality ✅ | Author Authority ✅ | Actionability ✅