TypeScript surpasses Python to become the most widely used language on GitHub, with AI being the main driving force.

What is the most commonly used programming language among developers? I believe many people would choose Python without hesitation.

However, the "Octoverse 2025" report recently released by GitHub gives a different answer: TypeScript.

According to GitHub's statistics on the number of contributors, in August 2025, after leading the list for 16 consecutive months, the number of Python contributors dropped to second place for the first time. TypeScript became the most widely used language on GitHub for the first time, surpassing Python with an advantage of approximately 42,000 contributors. JavaScript followed closely, and Java, C#, and PHP ranked fourth, fifth, and sixth with extremely stable positions.

The GitHub report states: "This milestone event is a concentrated manifestation of the trend of developers shifting towards typed JavaScript over the past decade and also marks that TypeScript is becoming the new default option for modern development."

More specifically, the number of TypeScript contributors increased by over 1 million in 2025 (a year-on-year increase of 66%). Its main driving forces come from development frameworks that default to using TypeScript for project setup on the one hand, and AI-assisted development on the other hand, as TypeScript's stricter type system benefits AI-assisted development greatly.

However, GitHub also points out that Python still maintains its dominant position in the fields of AI and data science, with 2.6 million contributors (a year-on-year increase of 48%). Jupyter Notebook remains the preferred exploratory environment in the AI field (there are approximately 403,000 relevant repositories; a year-on-year increase of 17.8% in projects with the AI tag).

The number of JavaScript contributors is still large (2.15 million), but its growth has slowed down as developers gradually shift towards TypeScript.

All in all, TypeScript and Python currently have a total of over 5.2 million contributors (accounting for approximately 3% of all active developers on GitHub in August 2025). The rise of typed languages indicates that AI is not only changing the speed of coding but also influencing which languages development teams choose when "trusting and adopting AI-generated code into the production environment".

In addition, according to GitHub statistics, 80% of the newly added software libraries in the past 12 months are concentrated in 6 core languages: Python, JavaScript, TypeScript, Java, C++, and C#.

Why did TypeScript win in 2025?

In August 2025, TypeScript ranked first on GitHub with 2,636,006 monthly contributors (a year-on-year increase of 1.05 million; +66.6%) and led in the number of newly added repositories.

The reason is the assistance of "types" to AI systems: The type system can reduce the ambiguity of code and catch errors of large language models (LLMs) early before the (AI-generated) code enters the production environment.

In addition, many frameworks also have TypeScript built - in by default. Next.js 15, Astro 3, SvelteKit 2, Qwik, SolidStart, Angular 18, and Remix all generate TypeScript codebases by default (through commands like npm create, pnpm dlx, or bunx create).

Moreover, the type system also helps to identify compilation errors generated by LLMs earlier in the development process. An academic study in 2025 found that 94% of the compilation errors generated by LLMs were due to failed type checks.

The entry threshold for TypeScript is also relatively low. Tools such as Vite, ts - node, Bun, and IDE auto - configuration hide the (cumbersome) boilerplate files, so junior developers can quickly start a typed technology stack.

Python still dominates AI projects

Even though TypeScript has risen, Python is still the well - deserved leader among all repositories with the AI tag. Among them, the usage of Jupyter Notebook almost doubled in 2025, which fully proves Python's status as the preferred language for AI workload prototyping, model training, and task orchestration.

Specifically, Python drives nearly half of the newly added AI repositories (582,196; a year-on-year increase of 50.7%), highlighting its role as the pillar for applied AI work (from training, inference to orchestration and deployment). Jupyter Notebook remains the preferred exploratory environment for experiments (402,643; a year-on-year increase of 17.8%), but the shift of (contributors) to Python codebases indicates that more projects are moving out of the prototype stage and entering the production technology stack.

Front - end and application - layer languages have achieved a sharp increase on a relatively small base: TypeScript increased by 77.9% (85,746) and JavaScript increased by 24.8% (88,023).

This indicates the rise of demos, dashboards, and lightweight applications built around model API endpoints.

Shell scripts (+324%) became the fastest - growing category, reflecting how teams are coding evaluation tools, data preparation, and deployment processes. C++ crossed 7,800 repositories (+11%), steadily reminding people of its role in performance - critical inference engines, runtimes, and hardware - close systems.

Other trends and highlights

GitHub also summarized some other trends and highlights in the report.

This year, open - source development activities reached a record level. The total number of contributions to public repositories reached 1.12 billion (a year-on-year increase of 13%). March 2025 was the month with the largest number of newly added open - source contributors in GitHub's history.

India has witnessed rapid growth. The country added over 5.2 million developers in 2025, accounting for more than 14% of the total 36 million newly added developers on GitHub in 2025. This makes India the largest single source of newly added developers on GitHub this year, continuing its rapid growth momentum since 2020.

Enterprise - level technology stacks remain stable. The number of contributors to Java and C# both increased by over 100,000 this year, indicating that even though AI is reshaping the entire industry landscape, their growth in large enterprises and game - development environments remains stable.

Experimental projects of old languages are emerging. The COBOL language also appears in GitHub's dataset, with nearly 3,000 active developers. This is probably driven by some organizations and enthusiasts who have created many AI - assisted tutorial repositories to help modernize legacy codebases.

Performance and system languages are rising with AI (but with uneven growth). The C language increased by approximately 20.9% year - on - year, and C++ increased by approximately 11.8% year - on - year, reflecting the market's demand for faster runtimes, inference engines, and hardware - optimized loops.

Generative AI is increasingly becoming infrastructure. Now, over 1.1 million public repositories have imported LLM SDKs (a year - on - year increase of 178% when comparing August 2025 with August 2024), supported by over 1.05 million contributors, and the monthly number of commits reaches 1.75 million (a 4.8 - fold increase since 2023).

The application of AI in the open - source field. Half (50%) of open - source projects have at least one maintainer using GitHub Copilot.

.NET remains strong. C# increased by approximately 10.6% year - on - year, consistent with the growth of enterprise - level and game/tool ecosystems. This indicates that AI features are being integrated into existing.NET workflows rather than driving a complete language shift.

The fastest - growing language is Luau. Luau is the scripting language of Roblox and a gradually typed language, reflecting the industry's trend towards "type flexibility".

Reproducibility and dependency hygiene are highly concerned. The rise of astral - sh/uv and NixOS/nixpkgs indicates developers' desire for deterministic builds, faster installation speeds, and direct operation.

Performance - centered developer tools have attracted attention. The common features of Ghostty, Tailwind CSS, and uv are all about speed, compact feedback loops, and minimized development resistance.

For more details, please refer to the original report:

https://github.blog/news-insights/octoverse/octoverse-a-new-developer-joins-github-every-second-as-ai-leads-typescript-to-1/

This article is from the WeChat official account "Almost Human" (ID: almosthuman2014), author: Panda, published by 36Kr with authorization.