84% of developers now use or plan to use AI coding tools in their workflows, yet only 29% trust the accuracy of what those tools produce, according to Stack Overflow’s 2025 Developer Survey of more than 49,000 respondents. That widening gap between use and trust defines the current state of AI-assisted software development, and the AI coding statistics below map how wide it has grown.
The data below maps adoption rates, tool market share, productivity gains, code quality, enterprise rollout, and the programming-language shifts that AI coding has set in motion.
Key Takeaways
- Over 84% of developers use or plan to use AI tools, up from 76% in 2024, per Stack Overflow.
- More developers distrust AI output (46%) than trust it (33%), with only 29% trusting accuracy.
- GitHub Copilot reached over 4.7 million paid subscribers, up 75% year-over-year, per Microsoft.
- Nearly 9 out of 10 developers save an hour weekly with AI tools, and 1 in 5 save eight-plus hours, per JetBrains.
- AI adoption lifts individual productivity but lowers delivery stability, per the 2024 DORA report of over 39,000 professionals.
- TypeScript became GitHub’s most-used language in August 2025, gaining roughly 1,054,015 contributors (+66.63%).
- Paid Microsoft 365 Copilot seats surpassed 20 million in Q3 FY2026, up 250% from a year earlier.
Editor’s Choice
- GitHub Copilot now contributes an average of 46% of code completed in enabled files, per GitHub’s economic-impact research.
- Developers completed tasks 55% faster with GitHub Copilot in quantitative studies.
- GitHub estimates a potential $1.5 trillion boost to global GDP from AI developer tools by 2030.
- 36.2 million new developers joined GitHub in 2025, more than one every second, bringing the platform total to over 180 million.
- More than 1.1 million public repositories now import an LLM SDK, up 178% annually.
- GitHub Copilot enterprise adoption reached nearly 140,000 organizations, tripled from a year earlier, by Q3 FY2026.
- JetBrains found 90% of developers regularly use at least one AI tool at work as of April 2026.
Recent Developments
- April 2026: JetBrains’ workplace survey found GitHub Copilot leads at 29% usage, ChatGPT at 28%, with Claude Code and Cursor tied at 18% each.
- April 2026: Microsoft disclosed GitHub Copilot enterprise adoption reached nearly 140,000 organizations on its Q3 FY2026 earnings call, tripled year-over-year.
- March 2026: Paid Microsoft 365 Copilot seats surpassed 20 million in Q3 FY2026, a 250% year-over-year increase, per Microsoft’s Form 10-Q.
- January 2026: GitHub Copilot crossed over 4.7 million paid subscribers, up 75% year-over-year, on Microsoft’s Q2 FY2026 earnings call.
- December 2025: Stack Overflow published its 2025 Developer Survey, showing AI adoption at 84% while trust in accuracy sat at just 29%.
AI Coding Statistics: Developer Tool Adoption
- Over 84% of developers use or plan to use AI tools in development, up from 76% in 2024, per Stack Overflow’s 2025 survey.
- 51% of professional developers use AI tools daily, per the same Stack Overflow data.
- 85% of developers regularly use AI tools for coding, per JetBrains’ State of Developer Ecosystem 2025.
- 62% of developers rely on at least one AI coding assistant, agent, or code editor, per JetBrains.
- 90% of developers regularly use at least one AI tool at work, per JetBrains’ April 2026 survey.
- 74% of developers have adopted specialized AI tools beyond general chatbots, per JetBrains’ April 2026 survey.
| Survey | Adoption Metric | Figure |
|---|---|---|
| Stack Overflow 2025 | Use or plan to use AI tools | 84% |
| Stack Overflow 2025 | Daily use (professionals) | 51% |
| JetBrains 2025 | Regularly use AI tools | 85% |
| JetBrains April 2026 | At least one AI tool at work | 90% |
| JetBrains April 2026 | Specialized AI dev tools | 74% |
Source: Stack Overflow 2025 Developer Survey, JetBrains Research
The 2025 Stack Overflow Developer Survey found that 84% of respondents use or plan to use AI tools in development, up from 76% in 2024, with 51% of professional developers using AI tools daily. JetBrains’ State of Developer Ecosystem 2025 found that 85% of developers regularly use AI tools for coding and development, while 62% rely on at least one AI coding assistant, agent, or code editor.
A separate JetBrains April 2026 survey found that 90% of developers regularly use at least one AI tool at work, with 74% having adopted specialized AI tools for developers beyond general chatbots.
The three surveys measure slightly different populations, yet they converge on the same conclusion: regular AI use has become the default developer behavior across the profession.
GitHub Copilot User and Revenue Statistics
- GitHub Copilot reached over 4.7 million paid subscribers, up 75% year-over-year, by Microsoft’s Q2 FY2026 earnings call.
- Copilot Pro Plus subscriptions for individual developers increased 77% quarter-over-quarter, per Microsoft.
- GitHub Copilot enterprise adoption reached nearly 140,000 organizations, tripled year-over-year, by Q3 FY2026.
- GitHub Copilot CLI usage almost doubled month-over-month, per Microsoft’s Q3 FY2026 call.
- Over 1 million developers have activated GitHub Copilot, generating over 3 billion accepted lines of code, per GitHub research.
- 80% of new GitHub developers used Copilot within their first week in 2025, per the Octoverse report.
| Metric | Figure | Period |
|---|---|---|
| Paid subscribers | 4.7 million (+75% YoY) | Q2 FY2026 |
| Pro Plus QoQ growth | +77% | Q2 FY2026 |
| Enterprise organizations | ~140,000 (3x YoY) | Q3 FY2026 |
| Activated developers | 1 million+ | Cumulative |
| Accepted lines of code | 3 billion+ | Cumulative |
Source: Microsoft FY2026 Earnings Calls, GitHub Research
GitHub Copilot reached over 4.7 million paid Copilot subscribers, up 75% year-over-year, as reported during Microsoft’s Fiscal Year 2026 Second Quarter earnings conference call, with Copilot Pro Plus subscriptions for individual developers increasing 77% quarter-over-quarter. GitHub Copilot enterprise adoption reached nearly 140,000 organizations, tripled year over year, with GitHub Copilot CLI usage almost doubling month over month, as disclosed on Microsoft’s Q3 FY2026 earnings call.
Over 1 million developers have activated GitHub Copilot, generating over 3 billion accepted lines of code, per GitHub’s research. 80% of new developers used Copilot within their first week in 2025, per the GitHub Octoverse report.
For the security profile of the code these tools generate, the AI-generated code vulnerability data tracks how often AI suggestions introduce exploitable flaws.
AI Coding Tool Market Share at Work
- GitHub Copilot leads workplace usage at 29%, with 76% awareness, per JetBrains’ April 2026 survey.
- ChatGPT follows at 28% workplace usage among developers.
- Claude Code and Cursor are tied at 18% workplace usage each.
- Copilot usage rises to 40% in companies with 5,000 or more employees.
- Claude Code surged roughly 6 times from approximately 3% workplace usage between April and June 2025.
- Among AI agent users specifically, ChatGPT leads at 81.7% and GitHub Copilot at 67.9%, per Stack Overflow.
| Tool | Workplace Usage | Source |
|---|---|---|
| GitHub Copilot | 29% | JetBrains April 2026 |
| ChatGPT | 28% | JetBrains April 2026 |
| Claude Code | 18% | JetBrains April 2026 |
| Cursor | 18% | JetBrains April 2026 |
| ChatGPT (agent users) | 81.7% | Stack Overflow 2025 |
| GitHub Copilot (agent users) | 67.9% | Stack Overflow 2025 |
Source: JetBrains Research, Stack Overflow 2025 Developer Survey
JetBrains’ April 2026 survey found GitHub Copilot leads workplace usage at 29% (with 76% awareness, and 40% in companies with 5,000 or more employees), followed by ChatGPT at 28%, and Claude Code and Cursor tied at 18% each. Claude Code surged roughly 6 times from approximately 3% workplace usage in April to June 2025. Among AI agent users, ChatGPT leads at 81.7% and GitHub Copilot at 67.9%, per the 2025 Stack Overflow survey.
By the numbers: Per JetBrains’ April 2026 workplace survey, GitHub Copilot (29%), ChatGPT (28%), Claude Code (18%), and Cursor (18%) split the developer market four ways. Claude Code’s roughly sixfold climb in under a year shows how quickly tool preference shifts in this category.
Developer Productivity Statistics with AI Coding Tools
- Developers completed tasks 55% faster with GitHub Copilot in quantitative research studies.
- Nearly 9 out of 10 developers save at least one hour weekly using AI tools, per JetBrains.
- 1 in 5 developers save eight or more hours per week, per JetBrains’ 2025 ecosystem report.
- Pull request merge rate rose 15% in the Accenture enterprise study.
- 95% of developers reported enjoying coding more with Copilot, per the Accenture study.
- 90% of developers expressed feeling more fulfilled with their jobs using Copilot.
| Productivity Metric | Figure | Source |
|---|---|---|
| Task completion speed | 55% faster | GitHub research |
| Developers saving 1+ hour/week | ~9 in 10 | JetBrains 2025 |
| Developers saving 8+ hours/week | 1 in 5 | JetBrains 2025 |
| Pull request merge rate increase | +15% | Accenture study |
| Enjoy coding more | 95% | Accenture study |
Source: GitHub research; JetBrains 2025; Accenture study
Developers completed tasks 55% faster with GitHub Copilot in quantitative research studies, per GitHub’s economic-impact research. Nearly 9 out of 10 developers save at least one hour weekly using AI tools, with 1 in 5 developers saving eight or more hours per week, per JetBrains. In the Accenture enterprise study, the pull request merge rate rose 15%, 95% of developers reported enjoying coding more, and 90% expressed feeling more fulfilled with their jobs.
The time-savings spread is the AI coding statistics figure enterprise buyers should watch. A tool that gives most developers an hour back but a fifth of them a full workday signals uneven returns that depend heavily on task type and team workflow.
Share of Code Written by AI Statistics
- GitHub Copilot contributes an average of 46% of code completed in enabled files, per GitHub research.
- Developers accepted around 30% of GitHub Copilot’s suggestions in the Accenture study.
- Developers retained 88% of Copilot-generated characters in final submissions.
- 91% of developers’ teams merged pull requests containing Copilot-generated code.
- 81.4% of developers installed the Copilot extension on day one of receiving their license.
- Over 3 billion accepted lines of code have been generated through Copilot, per GitHub.
GitHub Copilot now contributes an average of 46% of code completed in enabled files, per GitHub’s economic-impact research. Developers accepted around 30% of GitHub Copilot’s suggestions and retained 88% of Copilot-generated characters in their editor in final submissions, with 91% of developers’ teams merging pull requests containing Copilot-generated code and 81.4% installing the extension on day one.
The share of merged code that originates from AI suggestions raises governance questions that overlap with shadow AI usage statistics, where unsanctioned tool use complicates code-provenance tracking.
Developer Trust and Satisfaction with AI Coding Tools
- Only 29% of developers trust AI output accuracy, per Stack Overflow’s 2025 survey.
- 46% of developers actively distrust AI output, more than the 33% who trust it.
- Favorable sentiment toward AI dropped to 60% from 72% year-over-year.
- 66% of developers struggle with AI solutions that are almost right but not quite.
- 45% of developers find debugging AI-generated code time-consuming.
- 44% of developers used AI-enabled tools to learn new coding techniques, up from 37% in 2024.
Only 29% of respondents said they trust AI output accuracy, while 46% actively distrust it, with favorable sentiment declining to 60% from 72% year-over-year, per the 2025 Stack Overflow survey. A common frustration is AI output that is almost right but not quite, alongside time-consuming debugging.
45% of developers find debugging AI-generated code time-consuming, while a core frustration is AI solutions that are almost right but not quite, per Stack Overflow’s 2025 survey results. 44% of developers used AI-enabled tools to learn new coding techniques, up from 37% in 2024.
Key finding: Stack Overflow’s 2025 survey found favorable sentiment toward AI fell to 60% from 72% year-over-year, even as adoption rose to 84%. The “almost right but not quite” problem, the most-cited frustration, explains the divergence: developers use the tools constantly while trusting them less than before.
The trust gap mirrors a broader labor-market anxiety captured in AI job impact data, where developer sentiment about automation tracks closely with how much code AI now writes.
Enterprise AI Coding Adoption Statistics
- GitHub Copilot enterprise adoption reached nearly 140,000 organizations, tripled year-over-year, by Q3 FY2026.
- Paid Microsoft 365 Copilot seats surpassed 20 million in Q3 FY2026, a 250% year-over-year increase.
- Siemens deployed GitHub Copilot to 30,000 of its developers before adopting the full GitHub platform.
- Over 20,000 organizations have adopted Copilot for Business, per GitHub research.
- Copilot workplace usage rises to 40% in companies with 5,000 or more employees, per JetBrains.
- GitHub Copilot CLI usage almost doubled month-over-month, per Microsoft’s Q3 FY2026 call.
| Enterprise Metric | Figure | Source |
|---|---|---|
| Copilot enterprise organizations | ~140,000 (3x YoY) | Microsoft Q3 FY2026 |
| Microsoft 365 Copilot seats | 20 million (+250% YoY) | Microsoft 10-Q |
| Siemens Copilot deployment | 30,000 developers | Microsoft Q2 FY2026 |
| Copilot for Business orgs | 20,000+ | GitHub research |
| Copilot use in 5,000+ employee firms | 40% | JetBrains April 2026 |
Source: Microsoft FY2026 Earnings, GitHub Research, JetBrains
Paid Microsoft 365 Copilot seats surpassed 20 million in Q3 FY2026, representing a 250% increase year over year, reflecting rapid enterprise adoption across Microsoft’s commercial customer base. Siemens adopted the full GitHub platform after a successful Copilot rollout to 30,000 of its developers, per Microsoft’s Q2 FY2026 earnings call. Over 20,000 organizations have adopted Copilot for Business, per GitHub’s research. Copilot workplace usage rises to 40% in companies with 5,000 or more employees, per JetBrains’ April 2026 survey.
The enterprise rollout pace tracks alongside Microsoft 365 statistics, where Copilot seat growth is one of the fastest-expanding lines in Microsoft’s commercial business.
DORA 2024: AI Productivity vs Delivery Stability
- 75.9% of developers use AI for daily tasks such as writing code, summarizing, debugging, and testing, per the 2024 DORA report.
- The 2024 DORA report drew on insights from over 39,000 professionals worldwide.
- AI adoption increases individual productivity, flow, and job satisfaction, per DORA.
- AI adoption negatively affects software delivery stability and throughput, per the same report.
- DORA emphasizes that small batch sizes and robust testing remain crucial alongside AI adoption.
| DORA 2024 Finding | Direction |
|---|---|
| Daily AI task usage | 75.9% |
| Individual productivity | Increases |
| Flow and job satisfaction | Increases |
| Software delivery stability | Decreases |
| Delivery throughput | Decreases |
Source: DORA Accelerate State of DevOps Report 2024
The 2024 DORA State of DevOps Report, drawing from over 39,000 professionals worldwide, found that 75.9% of developers use AI for daily tasks such as writing code, summarizing information, debugging, and testing. The report identified a paradox: AI adoption significantly increases individual productivity, flow, and job satisfaction, but also negatively impacts software delivery stability and throughput, emphasizing that fundamentals like small batch sizes and robust testing remain crucial.
Worth noting: The 2024 DORA report of over 39,000 professionals found AI lifts individual productivity while it lowers delivery stability and throughput. That tension is the single most important caveat for engineering leaders: faster individual output does not automatically translate into faster, safer shipping.
The DORA finding cuts against the simplest sales pitch for AI coding tools. Productivity at the keyboard is not the same as predictable delivery, and the gap is exactly where governance and testing discipline earn their keep.
AI Code Quality Statistics
- A controlled trial assigned 104 developers Copilot access and 98 to a control group, all with at least five years of experience.
- Code written with Copilot showed readability improved by 3.62% and reliability by 2.94%.
- Maintainability improved 2.47% and conciseness 4.16% in the same trial.
- Copilot users showed a 53.2% greater likelihood of passing all 10 unit tests.
- Expert reviewers conducted 1,293 code evaluations using standardized rubrics.
- Copilot users averaged 18.2 lines of code per error, compared to 16.0 without assistance.
| Code Quality Dimension | Improvement | Significance |
|---|---|---|
| Readability | +3.62% | p=0.003 |
| Reliability | +2.94% | p=0.01 |
| Maintainability | +2.47% | p=0.041 |
| Conciseness | +4.16% | p=0.002 |
| Unit test pass likelihood | +53.2% | (greater likelihood) |
Source: GitHub Copilot Code Quality Controlled Study
In a controlled trial involving 202 qualified developers with a minimum of five years of experience, code written with GitHub Copilot showed readability improved by 3.62%, reliability by 2.94%, maintainability by 2.47%, and conciseness by 4.16%. Developers using Copilot demonstrated a 53.2% greater likelihood of passing all 10 unit tests, and across 1,293 expert code evaluations, Copilot users averaged 18.2 lines of code per code error compared to 16.0 without assistance.
The quality gains are real but modest, in the low single digits for most dimensions. The standout is the unit-test pass rate, which suggests AI assistance helps most where the task has a clear correctness target rather than open-ended design.
GitHub AI Project and Repository Growth Statistics
- More than 1.1 million public repositories now import an LLM SDK, up 178% year-over-year.
- 36.2 million new developers joined GitHub in 2025, more than one every second, reaching over 180 million total.
- GitHub saw a 59% surge in contributions to generative AI projects in 2024.
- The number of AI projects overall rose 98% in 2024, per the Octoverse report.
- GitHub recorded nearly 150,000 total generative AI projects, with over 70,000 new public ones created in 2024.
| Growth Metric | Figure | Year |
|---|---|---|
| Repos importing LLM SDKs | 1.1 million+ (+178%) | 2025 |
| New developers | 36.2 million | 2025 |
| Total GitHub developers | 180 million+ | 2025 |
| GenAI contribution surge | +59% | 2024 |
| GenAI project growth | +98% | 2024 |
Source: GitHub Octoverse 2024 and 2025
More than 1.1 million public repositories now import an LLM SDK, up 178% year over year, and 36.2 million new developers joined GitHub in 2025, more than one new developer joining every second on average, bringing the total to over 180 million developers.
In 2024, there was a 59% surge in contributions to generative AI projects on GitHub and a 98% increase in the number of projects overall, with nearly 150,000 total generative AI projects and over 70,000 new public generative AI projects created in 2024 alone.
The agentic layer of this growth, where AI moves beyond code completion into autonomous tasks, is detailed in AI agent adoption data.
AI Coding and Programming Language Statistics
- TypeScript overtook Python and JavaScript in August 2025 to become GitHub’s most-used language.
- TypeScript gained roughly 1,054,015 contributors, a 66.63% year-over-year increase.
- TypeScript repositories in new projects grew 78.10% year-over-year as AI accelerated typed-language adoption.
- More than 1.1 million public repositories import an LLM SDK, up 178% year-over-year as of August 2025.
- A 2025 academic finding showed most LLM-generated compilation errors were type-check failures.
| Language Metric | Figure | Period |
|---|---|---|
| TypeScript contributor growth | +66.63% (~1,054,015) | YoY to Aug 2025 |
| TypeScript new-project repo growth | +78.10% | YoY |
| LLM SDK repo imports | 1.1 million+ (+178%) | As of Aug 2025 |
Source: GitHub Octoverse 2025
TypeScript overtook both Python and JavaScript in August 2025 to become the most used language on GitHub, gaining approximately 1,054,015 contributors, representing a 66.63% increase year over year. TypeScript repositories in new projects grew 78.10% year over year as AI-assisted coding accelerated adoption of typed languages, with a 2025 academic finding that most LLM-generated compilation errors were type-check failures, making TypeScript a natural companion to AI coding tools.
The TypeScript surge is the clearest structural fingerprint AI coding has left on the ecosystem. When LLMs produce type-check errors that a typed language catches at compile time, developers gravitate toward the language that absorbs the failure mode, and the adoption data follows.
The same open-source contribution trends that lifted TypeScript also show up in the Linux developer statistics, where platform-level shifts track language adoption closely.
AI Coding Economic Impact Statistics
- GitHub estimates a potential $1.5 trillion boost to global GDP from AI developer tools by 2030.
- That GDP boost is equivalent to adding 15 million effective developers to worldwide capacity.
- The estimate is based on 45 million projected professional developers in 2030.
- GitHub Copilot has generated over 3 billion accepted lines of code to date.
- Developers completed tasks 55% faster with Copilot in quantitative studies.
| Economic Metric | Figure | Horizon |
|---|---|---|
| Potential global GDP boost | $1.5 trillion | By 2030 |
| Effective developers added | 15 million | By 2030 |
| Projected professional developers | 45 million | 2030 |
| Accepted lines of code | 3 billion+ | Cumulative |
Source: GitHub Economic Impact Research
GitHub estimates a potential $1.5 trillion boost to global GDP from AI developer tools by 2030, equivalent to adding 15 million effective developers to worldwide capacity based on 45 million projected professional developers in 2030.
For developers building the skills behind these gains, demand for structured training on major learning platforms has climbed in parallel with tool adoption.
How Many Developers Use AI Coding Tools?
Most developers now use AI coding tools regularly, though the exact figure depends on the survey and population. Stack Overflow’s 2025 survey put adoption at 84% of respondents using or planning to use AI tools. JetBrains found 85% of developers regularly use AI tools for coding.
The 2024 DORA report measured 75.9% using AI for daily tasks. The spread across these surveys reflects different question framing, while all three land in the same high-adoption band.
Which AI Coding Tool Has the Most Users?
GitHub Copilot leads by paid subscribers, while ChatGPT leads among AI agent users. GitHub Copilot reached over 4.7 million paid subscribers, up 75% year-over-year, by Microsoft’s Q2 FY2026 call. By workplace usage, JetBrains’ April 2026 survey put GitHub Copilot first at 29%, with ChatGPT at 28%.
Among AI agent users specifically, ChatGPT leads at 81.7% versus Copilot’s 67.9%, per Stack Overflow. The answer depends on whether the metric is paid subscriptions, workplace usage, or agent-specific adoption.
Conclusion
The defining AI coding statistics this year remain the gap between use and trust: 84% of developers use AI tools while only 29% trust their accuracy, a divide that survey after survey shows widening each year. GitHub Copilot reached over 4.7 million paid subscribers, and paid Microsoft 365 Copilot seats surpassed 20 million in Q3 FY2026, confirming that adoption is no longer the open question. The unresolved question is whether the tools earn the trust their usage already assumes.
The DORA stability paradox and TypeScript’s rise to GitHub’s most-used language point to where the next phase plays out: governance that keeps delivery stable as individual output rises, and typed languages that absorb the errors AI still produces. The platform-level view behind these tool numbers, from repository growth to contributor counts, sits in the broader GitHub platform statistics that this data refreshes against each quarter.
