Sonar survey: AI now writes 65% of committed code, and 76% of developers can't verify it fast enough

AI now generates or significantly assists an average of 65% of the code developers commit, according to a survey of 1,026 developers and engineering managers by code quality company Sonar, described in a press release.

Checking that code is not keeping up. 76% of respondents said they generate AI code faster than they can verify it, and 53% now spend more time reviewing code than writing it, up from 18% before they adopted AI tools. Only 39% fully agree that AI-generated code is always checked before it is committed, although 94% say they do not fully trust it. Just 23% have clear, AI-specific guidelines or automated checks for AI-written code.

The stakes are rising: 76% say their teams use AI coding tools on business-critical or mission-critical services, and 34% expect to build or use "agentic software factories" within a year, up from 5% today. Sonar sells tools for verifying and governing code, so it has a stake in the review problem the survey describes.