JetBrains' research team conducts developer surveys every few months, and the latest data offers some interesting insights.
In January 2026, they surveyed over 10,000 professional developers worldwide on one question: which AI tools are you actually using at work?
The headline: 90% of developers are already using AI tools. But which tools are growing and which are stagnating tells a very different story.
GitHub Copilot: The veteran, but growth has stalled
GitHub Copilot is the most well-known AI coding tool, with 76% of developers having heard of it. However, only 29% actually use it at work, showing no significant growth compared to last year.
In large enterprises (5,000+ employees), Copilot remains dominant with a 40% adoption rate. This suggests it survives on enterprise compliance advantages and IT procurement inertia, but its appeal among engineers making independent choices is declining.
Claude Code: From 3% to 18%, a sixfold increase in six months
The most striking growth curve belongs to Claude Code:
| Period | Claude Code Adoption Rate |
| April–June 2025 | ~3% |
| January 2026 | 18% |
In six months, adoption grew sixfold. In North America, the figure is even higher at 24%.
Satisfaction metrics are equally impressive: CSAT (Customer Satisfaction) at 91%, NPS (Net Promoter Score) at 54. In enterprise software, these numbers are considered top-tier.
What's driving Claude Code? Survey data points to "performance over platform" — developers are willing to migrate to more effective tools regardless of ecosystem.
Cursor: Neck and neck with Claude Code
Cursor has 69% awareness and matches Claude Code at 18% workplace adoption.
When it comes to "primary tool" choice (the one used most often), Claude Code takes 28%, Cursor takes 24%, and together they account for over 50%.
Chat interfaces aren't dead yet
Interestingly, developers also use a fair number of chat-based tools for coding:
| Tool | Percentage Used for Development Tasks |
| ChatGPT | 28% |
| Gemini | 8% |
| Claude Chat Interface | 7% |
Many developers haven't yet moved from "asking AI questions" to "letting AI write code." Both usage patterns coexist without excluding each other.
Google Antigravity: Entering the fray
Google launched Antigravity (an AI tool for developers) in November last year. By January, 6% of developers were already using it. For a product just two months old, that's a decent start. Whether it can hold ground depends on future product strength.
Who will win the next round?
The underlying logic revealed by the survey is simple: Developers will actively migrate to better tools, and ecosystem lock-in is less powerful than before.
GitHub Copilot captured the market by being two years early and leveraging GitHub integration, but its growth has stalled. Claude Code, a latecomer, has caught up through product strength.
If this pattern continues, the future belongs to whoever has the stronger model and better tool experience.
Source: CocoLoop, Which AI Coding Tools Do Developers Actually Use at Work? (JetBrains Research Blog)