Introduction
The 2025 edition of The State of WebDev AI offers a detailed snapshot of how developers are using AI today, which tools have gained the most traction over the past year, and what these trends suggest about the future of the industry.
In this summary, we’ve dug into the latest data to better understand the current landscape of artificial intelligence and where it may be heading.
Demographics
Before we get into the results, it’s worth saying that survey respondents were predominantly white (77%), male (90%), and non-disabled (73%).
While the United States accounted for the largest share (19%), there was also strong representation from European countries such as Germany, France, Poland and the United Kingdom (25% combined).
Tool Usage and Popularity
Among AI tools, unsurprisingly ChatGPT was by far the most widely used, with 91% of respondents having tried it, though this does not mean it’s the LLM they use most.
Around 55-60% of respondents have used Claude, Microsoft Copilot or Google Gemini - and 45% have used DeepSeek - with 9 out of 10 developers having heard of all four regardless of having used them. This shows product awareness is relatively strong, and even among the big players in the market.
xAI’s Grok appears less popular than many of its competitors, with only 25% of respondents having used it.
IDEs and Code Tools
When it comes to AI-driven IDEs, Cursor dominates the market, with over 80% of respondents having either used or heard of it, with Zed following at a distant second, with about 54% awareness.
For traditional coding environments, VSCode remains the clear leader, used by 41% of respondents. Other popular choices include GitHub Copilot (19%), Trace (15%), and Neovim (10%). Beyond these, usage drops off quickly across a range of smaller tools like JetBrains AI and Cline.
Code Generation Tools
In terms of code generation, v0 - backed by Vercel - has quickly taken the lead as the most popular tool, with 25% of all respondents having used it. Bolt from StackBlitz follows closely behind, maintaining solid awareness and usage numbers.
Other entrants like Lovable, and OpenUI show far lower usage rates, with a majority of respondents having never heard of them. This suggests the market is still very top-heavy, with a few early leaders establishing dominance.
Programming Languages and Image Generation
While Python is often closely tied to AI development, JavaScript/TypeScript were by far the most popular languages to use with AI. It’s no surprise (remember, it’s the State of WebDev AI) when you remember that web developers are mostly calling APIs using their favourite language, rather than training a model.
In image generation, DALL-E led by a wide margin, followed by Midjourney and Stable Diffusion. Notably, over 40% of respondents reported not using any AI image tools at all, showing this side of generation is still relatively niche among developers.
How Developers Are Using AI
The most common AI use case among developers is code generation, followed by learning and research and text generation. While image generation helped kick off the AI boom, only about 38% of respondents reported using it in practice.
When it comes to AI-assisted coding, the technology is still mainly a helper rather than a full automation tool. 60% of developers generate less than 25% of their code using AI, and only a small minority (8%) rely on AI for more than 75% of their output.
With the rise of vibe coding, it’s surprising to see that most developers only use it to write a relatively small fraction of the code they produce.
Pain Points
Coding assistants and code generation tools, the biggest pain points were hallucinations and inaccuracies, context and memory limitations, intrusive suggestions and poor code quality (13%).
Most developers can relate to the frustration of using a hallucinating AI while it writes poor quality code, or forgets what it did a few moments before and writes nonsensical code, so it’s interesting to see these popping up as pain points here.
Conclusion
Overall, The State of AI 2025 paints a picture of a fast-maturing ecosystem, where AI is firmly embedded in everyday development — but primarily as an enhancement, not a replacement.
While ChatGPT, Cursor, and v0 have established clear leads in their respective categories, the market remains dynamic, with emerging tools like Grok and Bolt gaining traction.
The report also underscores the fact that developers are, perhaps quite surprisingly, still writing the majority of the code they create rather than just vibecoding.
We’ll keep an eye on the State of AI and we’re already looking forward to seeing how it looks in 2026.
