Programming Language Job Market 2026 — Data Analysis

Ask which programming language is "best" and you get a different answer for each definition of the word. Most used, fastest growing, best paid, and easiest to learn point at four different languages. This analysis pulls together the public datasets that can be checked, namely the 2025 Stack Overflow Developer Survey, GitHub's Octoverse report, and US Bureau of Labor Statistics projections, and says plainly where the data stops. The short version: Python and TypeScript are where growth is concentrated, but the language you list on a résumé predicts less about your pay than your role and your country.

Why language "demand" is harder to measure than it looks

Counting how many people use a language is easy. Counting how many employers are paying for it is not, and the two numbers are often confused. Think of a highway. A traffic counter tells you how many cars pass a point, but not how many of those drivers are being paid to drive, and not how many trucks are hauling freight that someone is waiting for.

Three kinds of public evidence exist, and each answers a different question:

Source	What it measures	What it cannot tell you
Stack Overflow Developer Survey 2025	Self-reported use of each language in the past year, by respondents who chose to take the survey	Whether those respondents are hired for that language, or how many openings exist
GitHub Octoverse 2025	Distinct monthly contributors to repositories on GitHub (August 2025 snapshot)	Closed-source enterprise work, which does not show up on GitHub
US Bureau of Labor Statistics (BLS)	Employment, median pay, and 10-year projections for broad occupations	Anything at the level of a single language

Job-board counts are real, but they are not published as a citable, stable series. Postings also double count: one back-end job that asks for "Python, SQL, and AWS" appears under three languages. Anyone quoting precise posting counts by language, with a year-over-year percentage, should be asked where the series lives.

The sections below use only figures that appear on the cited pages, and mark clearly where the text is interpretation rather than measurement.

The 30-second answer

• Use: JavaScript (66%), HTML/CSS (61.9%), SQL (58.6%), Python (57.9%), and TypeScript (43.6%) are the five most-used languages among all 2025 Stack Overflow respondents. Stack Overflow
• Growth: On GitHub, TypeScript had 2,636,006 monthly contributors in August 2025, up 66.6% in a year, and passed both Python and JavaScript. GitHub Octoverse
• Jobs overall: BLS projects 10% growth in software developer, QA analyst, and tester employment from 2025 to 2035, with about 106,100 openings per year. BLS

What developers actually use: the 2025 Stack Overflow numbers

The Stack Overflow survey is the largest regular public snapshot of what working programmers touch. The table below reproduces its language usage figures for all respondents and for professional developers, which is the more relevant column for a job-market question.

Language	All respondents	Professional developers
JavaScript	66.0%	68.8%
HTML/CSS	61.9%	63.0%
SQL	58.6%	61.3%
Python	57.9%	54.8%
TypeScript	43.6%	48.8%
Java	29.4%	29.6%
C#	27.8%	29.9%
C++	23.5%	21.8%
C	22.0%	19.1%
PHP	18.9%	19.1%
Go	16.4%	17.4%
Rust	14.8%	14.5%
Kotlin	10.8%	11.5%
Ruby	6.4%	6.9%
Swift	5.4%	5.7%

Source: Stack Overflow Developer Survey 2025, technology section. Respondents could select multiple languages, so columns add to far more than 100%.

Two patterns stand out. First, the professional column and the all-respondents column tell different stories for the two headline languages. TypeScript climbs from 43.6% to 48.8% among professionals, while Python slips from 57.9% to 54.8%. In plain terms, TypeScript skews toward people who are paid to write code, and Python skews toward people who are learning, studying, or doing data work alongside another job.

Second, the survey itself flags Python as the biggest mover: it "saw a 7 percentage point increase from 2024 to 2025." A seven-point jump in a single year in a field this large is rare, and it lines up with the surge in AI and data work described below.

The survey also splits usage by whether respondents use AI tools. Among learners who use AI, 73.8% reported using Python and 35.1% reported TypeScript. Among professionals who use AI, the figures are 56.1% and 51.4%. The gap says something about the entry point into programming in 2026: newcomers are being steered, or are steering themselves, toward Python much more than toward the web stack.

Python: the default language of AI work

Python's position rests on one structural fact. The tools that people use to build, train, and run machine learning systems are overwhelmingly written to be called from Python. GitHub's Octoverse reports that Python is the leading language in AI-tagged repositories, with 582,196 repositories, up 50.7% year over year, and says it "now powers nearly half of all new AI repositories." Those are repository counts, not people, but they show where the new AI work is landing.

Python's lead is also a network effect. A hiring manager looking for someone to maintain a data pipeline, a notebook-based analysis, or a model-serving service will specify Python because the libraries, the tutorials, the existing code, and the available candidates are all there. The language becomes more useful because many people use it, and more people use it because it is useful. Economists call that a platform effect; for a learner it simply means more places to apply the same skill. Anyone starting from zero can follow how to learn Python in 2026, and readers weighing an AI-focused career can compare it with the ML engineer career path.

What Python is changing under the hood

Python's historical weak spot for performance-heavy and parallel work has been the Global Interpreter Lock (GIL), a mechanism that lets only one thread run Python bytecode at a time. Python 3.14, released on October 7, 2025, made the free-threaded build, which removes the GIL, officially supported under PEP 779. The release notes put the single-threaded penalty of that mode at "roughly 5-10%, depending on the platform and C compiler used." That is a trade-off rather than a free upgrade: code that runs one thread pays a small tax in exchange for the option of using many cores. How much of the ecosystem's compiled extensions will follow is a question the next few years will answer. Source: Python 3.14 release notes.

Tooling is also moving. Package and environment management, historically one of Python's annoyances, now has fast alternatives such as the Rust-built tool covered in What Is uv?. None of that changes the hiring picture directly, but it lowers the friction that pushes some teams to other languages.

Where Python does not win

Python is a poor first choice for several kinds of work, and the job market reflects that:

• Native mobile apps, which are written in Swift or Kotlin.
• Game engines and other latency-sensitive systems, where C++ and C# dominate.
• Firmware and embedded code, where C, C++, and increasingly Rust are standard.
• Browser code, which is JavaScript or TypeScript regardless of what runs on the server.

A Python job is therefore usually a back-end, data, automation, or machine learning job. Within that space, web back-ends are often built with one of the frameworks compared in FastAPI vs Flask vs Django, and model work leans on the libraries in PyTorch vs TensorFlow vs JAX.

TypeScript: the quiet number one on GitHub

The most striking public data point of the past year is not about Python at all. According to GitHub, TypeScript "overtook Python and JavaScript in August 2025 to become the most used language on GitHub," measured by distinct monthly contributors. The numbers from the report:

Language	Monthly contributors (August 2025)	Year-over-year change
TypeScript	2,636,006	+66.6% (about +1.05 million)
Python	about 2.6 million	+48.78% (about +850,579)
JavaScript	about 2.15 million	+24.79% (about +427,148)

Source: GitHub Octoverse 2025. TypeScript led Python by roughly 42,000 contributors. GitHub notes that other indices use different methods and may still rank JavaScript or Python first.

That caveat matters. A contributor count measures activity on one platform, and the ranking flips depending on whether you count commits, repositories, or people. Still, a 66.6% annual jump from an already large base is not noise.

Why TypeScript grew so fast

GitHub's own explanation has three parts. First, popular frameworks now generate TypeScript projects by default, citing Next.js, Astro, and SvelteKit. A developer who runs a "create new app" command and accepts the defaults is writing TypeScript without having decided to. Second, build tooling such as Vite and Bun has removed most of the configuration work that used to make TypeScript a chore to adopt. Third, the report argues that type systems suit AI-assisted coding: "Typed contracts scale AI-assisted teams." It cites a 2025 academic study finding that 94% of compilation errors in LLM-generated code were type-check failures, which a typed language catches before anything runs.

That third point is an argument from the platform vendor, and it should be read as a hypothesis rather than a proven cause. It is, however, consistent with the professional-versus-learner split in the Stack Overflow data. For a deeper look at the language itself, see TypeScript for Beginners, and for the framework layer that now defaults to it, Next.js vs React.

What happens to JavaScript

TypeScript compiles to JavaScript, so the two are not rivals in the way that Python and Java are. JavaScript remains the most-used language in the Stack Overflow survey, at 66% overall and 68.8% among professionals, because every browser runs it and because TypeScript code ships as JavaScript. The better framing is that "JavaScript" increasingly names the runtime and "TypeScript" names the language people write. GitHub described JavaScript's growth as slowing, but 24.79% annual growth in contributors is still large. A developer who knows TypeScript well is employable in almost every JavaScript role, while the reverse is becoming less true for new projects.

Java and C#: the enterprise middle

Java and C# sit near each other in the survey: Java at 29.4% of all respondents and 29.6% of professionals, C# at 27.8% and 29.9%. Both are statically typed, garbage-collected languages with large standard libraries, long-term support releases, and decades of production code behind them.

The honest summary of their job-market role is stability. Many large organizations run core systems on one or the other, and those systems need maintainers, upgrades, and new features for as long as the organization exists. GitHub's report shows both still growing, with Java adding about 174,705 monthly contributors (+20.73%) and C# about 136,735 (+22.22%) over the year. Those are healthy rates, though well below TypeScript's.

What the public data cannot say is how much of this work is hired through channels that never show up on GitHub or in a survey. In the 2025 data, a developer picking Java or C# is making a bet on a large, slow-moving, established employer base rather than on rapid growth. That is a perfectly reasonable bet. A bank's payments system does not get rewritten because a newer language is fashionable.

Go: small share, specific territory

Go was used by 16.4% of all survey respondents and 17.4% of professionals, which is a bigger footprint than Ruby, Swift, or Kotlin. It is a compiled language with a small feature set and built-in support for concurrency (running many tasks at once), designed to make server software easy to build and deploy as a single binary.

Its strongest niche is infrastructure and cloud tooling. Kubernetes, the container orchestration system that much of the industry runs on, is a Go project. Teams that build or extend that kind of software, or that need a fast, simple service language without the ceremony of Java, tend to reach for Go. A rough analogy is a Swiss Army knife with only three blades: it does fewer things than a full toolbox, but each of them is quick to pick up, and a new team member can read most of the code on day one.

The limit is breadth. Go is not a language for data science, mobile apps, or browser code, so a Go-focused career is a back-end and infrastructure career. That is a smaller pool of employers than Python or TypeScript serve, which cuts both ways: fewer openings, but also fewer candidates competing for each.

Rust: still small, no longer a side project

Rust was used by 14.8% of all respondents and 14.5% of professionals in the 2025 survey. That is close to Go in share, which is notable for a language that is much harder to learn. Rust enforces memory safety (preventing a class of bugs where a program reads or writes memory it should not) at compile time, without a garbage collector, which is why it appeals to people who would otherwise write C or C++.

The clearest evidence of its permanence is the Linux kernel. Rust support was merged in kernel version 6.1 as an experiment, to find out whether it was "worth the tradeoffs, technically, procedurally and socially." In December 2025, a patch by kernel developer Miguel Ojeda removed the "experiment" label from the kernel documentation, following the 2025 Maintainers Summit's conclusion that the experiment was over. The patch notes that Rust is used in production and ships in millions of Android devices, and it adds a caveat that matters: support is not uniform across every configuration, architecture, and toolchain, and substantial work remains. Source: LKML patch "rust: conclude the Rust experiment".

For career planning, the useful reading is narrow. Rust jobs exist mostly where C or C++ used to be the only option: systems software, security-sensitive components, performance-critical infrastructure. The survey shows many people using the language, but usage is not the same as paid openings. Treat it as a strong second language for a systems-minded developer, not as the default first one.

C and C++: durable, specialized

C (22.0%) and C++ (23.5%) together remain a large share of developer activity, and C++ is more common among all respondents than among professionals (23.5% vs 21.8%). These languages run operating systems, game engines, browsers, databases, embedded firmware, and the performance-critical cores of machine learning libraries. Their job market is specialized rather than broad: employers hire C++ developers for a particular kind of problem, and the skill transfers poorly to web or data roles.

The risk for a C++ specialist is not decline in the near term, because the installed base is enormous. It is narrowness. A developer who can write correct low-level code in a language without a safety net has a hard-to-replace skill, but the number of employers that need it is smaller than for general-purpose web or data work.

Mobile: Kotlin and Swift

Kotlin was used by 10.8% of respondents and 11.5% of professionals, and Swift by 5.4% and 5.7%. Google's Android documentation says Kotlin is "used by over 60% of professional Android developers" and describes Android Studio as providing first-class support for it. That is the vendor's claim and should be read as such, but it is consistent with Kotlin's position as the main Android language. Swift's smaller share reflects that iOS development is a single platform with one primary language.

Both languages also face competition from cross-platform approaches. Kotlin Multiplatform lets teams share code between Android, iOS, and desktop. React Native, which is written in JavaScript or TypeScript, is another route, so a web developer sometimes ends up in a mobile job without learning Swift or Kotlin at all. Public data do not say how the mobile market splits between native and cross-platform work.

PHP and Ruby: smaller shares, long tails

PHP was used by 18.9% of all respondents and 19.1% of professionals, and Ruby by 6.4% and 6.9%. The survey page does not show a trend for either, so this article does not claim that they are growing or shrinking. What the numbers do show is that PHP is still a mainstream language by share, ahead of Go and Rust, which matches the reality that a large portion of the web runs on content management systems and legacy applications built in it. Ruby's share is much smaller, and its hiring is concentrated in organizations with long-lived Rails applications.

For a career decision, the useful question is not whether a language is "dying" but who pays for it. Maintenance work on large existing systems can be steady and well paid even when new projects choose something else, though it tends to offer less room to grow into new kinds of work.

What the pay data actually shows

Searching for "highest paying programming language" produces confident rankings, almost all of which rest on self-selected samples. The Stack Overflow 2025 work section reports median annual compensation by developer role and by country, not by language. Here are the figures it gives for developer roles:

Role	United States	Global	Germany	United Kingdom
Developer, back-end	$175,000	$79,742	$87,011	$108,913
Developer, front-end	$145,000	$62,015	$79,636.50	$84,544
Developer, full-stack	$138,000	$72,509	$75,410	$85,428.50
Developer, QA or test	not listed	$57,442.50	not listed	not listed

Source: Stack Overflow Developer Survey 2025, work section. Medians in US dollars, self-reported.

The first thing the table shows is the size of the geography effect. A US back-end developer's reported median of $175,000 is more than twice the global median for the same role ($79,742) and roughly twice that of Germany. Choosing a country moves compensation by much more than choosing between any two mainstream languages. India's back-end median, listed on the same page, is $22,086.

The second is that role matters more than the language label. Back-end developers out-earn front-end developers in the US figures by $30,000, regardless of whether the back end is written in Java, Python, or Go. Roles that overlap heavily with particular languages, such as machine learning engineer or low-level systems engineer, may carry premiums, but the public survey page does not break pay out by language, and claims that "Rust pays $230K" or "C++ pays $250K+" should be traced to a named dataset before anyone relies on them.

The broader labor market

For the occupation as a whole, BLS reports a median annual wage of $134,040 for software developers, quality assurance analysts, and testers combined in 2025, with software developers alone at $135,980 and QA analysts and testers at $104,300. Employment was 1,905,400 in 2025 and is projected to grow 10% from 2025 to 2035, which BLS classes as much faster than average. About 106,100 openings are projected per year, and BLS notes that most come from workers who transfer to other occupations or leave the labor force rather than from new positions.

Put in proportion, 106,100 openings against 1,905,400 jobs is roughly 5.6% a year. That is the churn of an occupation, and it means a language skill is useful only if it connects to one of those openings. Source: BLS Occupational Outlook Handbook.

Does AI change which language to learn?

Two separate questions hide inside that one. The first is whether AI coding tools change what employers want. The second is whether they change who gets hired.

On the first, the evidence in this article is indirect. GitHub argues that typed languages suit AI-assisted work, and the Stack Overflow data show that Python is the dominant language among learners who use AI. Neither proves that AI assistance has shifted employer requirements, and it would be a mistake to build a career plan on the argument. A reasonable reading is that mainstream languages with large amounts of public code, like Python and TypeScript, are simply the ones that AI tools handle most fluently, because they have the most training material. For more on how these tools behave in practice, see What Is Vibe Coding?.

On the second, there is a published study. Stanford Digital Economy Lab researchers using ADP payroll data report that employment for workers aged 22 to 25 in AI-exposed occupations is about 19% below where it would be had it tracked less-exposed peers, in the current revision of "Canaries in the Coal Mine?", with data through June 2026. Experienced workers show no comparable gap. The authors find no evidence of broad economy-wide displacement, locate the effect in reduced hiring of young workers rather than increased separations, and describe their results as early and descriptive rather than causal. The page does not single out software developers by name. Source: Stanford Digital Economy Lab.

For someone choosing a first language, the takeaway is cautious. The data suggest entry-level hiring in exposed occupations may be harder than it was, which makes a visible portfolio and a second skill more important than the choice between two popular languages. Broader context sits in Will AI Take My Job? and The Future of Work.

Which language fits which career

The table below is interpretation, not measurement. It maps the usage and growth data above onto common career paths, with the reasoning shown so readers can disagree with it.

Career path	Languages that fit	Why
Machine learning or data work	Python, SQL	Python leads AI repositories on GitHub and is used by 57.9% of survey respondents; SQL is the third most-used language
Web front end or full-stack	TypeScript, JavaScript, HTML/CSS	TypeScript is the top language on GitHub by contributors and is defaulted to by major frameworks
Back-end at an established company	Java or C#, with SQL	Both are used by close to 30% of professionals and have long-term enterprise footprints
Cloud and infrastructure tooling	Go, Python	Kubernetes is a Go project; Python is widely used for automation and scripting
Systems, embedded, and game engines	C++, C, Rust	These replace each other more often than they replace higher-level languages
Android	Kotlin	Google says over 60% of professional Android developers use it
iOS	Swift	Single-platform primary language

For data-oriented readers, Python vs R for Data Science goes deeper on the first row, and the ML job market analysis and web development job market analysis cover the first two rows from the hiring side.

Where the data is thin, and what could change

A few honest limits apply to everything above.

1. Survey self-selection. Stack Overflow respondents are people who found and chose to complete a long survey. They skew toward developers who use the site, and the survey's own all-respondents and professional columns differ by several points for the top languages.
2. Platform bias. GitHub measures open source and the private work of teams that host there. Companies with on-premises systems or other hosts are underrepresented, and GitHub itself notes that other indices rank languages differently.
3. No language-level pay series. The best public sources report pay by role and location. Language is only one of the things that varies among developers.
4. Fast-moving tooling. Python's free-threaded builds, Rust in the Linux kernel, and TypeScript defaults in major frameworks are each less than two years old as established facts. Rankings based on 2025 data may look different within a year.

Three developments are worth watching, flagged as judgment rather than prediction. If Python's free-threaded mode proves stable across the extension ecosystem, some performance work that is currently written in C++ or Rust might stay in Python. If Rust's role in operating systems widens beyond Linux drivers, demand for it could grow faster than its current usage share suggests. And if entry-level hiring continues to soften, the market may reward breadth, such as a language plus testing, deployment, and data skills, more than depth in one syntax. Any of these could turn out wrong. For the international angle, countries with the most tech workers and countries with the most open source contributors show where the developer population is concentrated.

Common misunderstandings about language rankings

"The most popular language has the most jobs." Popularity in a survey counts people who use a language, including students and hobbyists. Python's 57.9% among all respondents drops to 54.8% among professionals, and TypeScript rises from 43.6% to 48.8%. The professional column is the better guide to what employers use.

"A newer language pays more." The public pay data does not support a language-level claim either way. It supports a role-level and country-level claim. A US back-end developer's median ($175,000) exceeds a front-end developer's ($145,000) by $30,000 in the survey, a gap that has nothing to do with which language is newer.

"TypeScript replaced JavaScript." TypeScript compiles to JavaScript, so every TypeScript application is also a JavaScript application when it runs. What changed is which language developers write, not what the browser executes.

"Learning the right language guarantees a job." BLS projects 106,100 openings a year in an occupation of 1,905,400 people, most of them from turnover. Competition for those openings depends on experience, portfolio, and location. A language is a tool that opens some doors, not a ticket.

FAQ

Q: Should I learn one language or several? A: The evidence in this article does not settle that, but it points toward a first language that matches a career path, followed by a second that complements it. Python plus SQL covers data work, and TypeScript plus HTML/CSS covers most web work. Depth in one language makes later ones faster to learn, and employers often want to see that you have built something non-trivial, not that you have touched many syntaxes.

Q: Is Python still the best first language in 2026? A: For people interested in data, automation, or AI, it is the language with the strongest evidence behind it. It rose seven percentage points in the 2025 Stack Overflow survey, and 73.8% of learners who use AI tools reported using it. For web development, TypeScript is a more direct route, because it is the language major frameworks now generate by default.

Q: Does Rust pay more than other languages? A: Public data cannot say. The Stack Overflow survey page reports pay by role and country rather than by language. Rust usage is 14.8% of respondents, and the language is now a permanent part of the Linux kernel, but neither fact translates into a salary figure. Claims about a Rust pay premium should be traced to a named dataset before being relied on.

Q: Which programming languages are fading? A: The 2025 survey page this article uses gives usage shares, not trends, for PHP (18.9%), Ruby (6.4%), and Swift (5.4%), so it does not show fading. These languages have smaller shares than the leaders, but they still support significant bodies of production code. A small or flat share does not mean there are no jobs.

Q: How much do software developers earn? A: BLS reports a 2025 median of $134,040 for software developers, QA analysts, and testers combined, with software developers at $135,980. Stack Overflow's self-reported US medians are higher for back-end ($175,000), front-end ($145,000), and full-stack ($138,000) developers, partly because its respondents are not a random sample of all programmers.

Q: Will AI coding tools reduce demand for programmers? A: Evidence so far is mixed and early. BLS projects 10% growth for the occupation through 2035. A Stanford study using payroll data finds employment of 22 to 25 year olds in AI-exposed occupations about 19% below trend, with no comparable gap for experienced workers, but its authors call the results descriptive rather than causal.

The takeaway

The cleanest finding in the 2025 data is a divergence between two kinds of popularity. Python is the language that more people are learning and using for AI work, while TypeScript is the one that professionals use more than the overall population and that now leads GitHub by contributors. Java, C#, C++, and Go keep large, specialized job markets that do not depend on being fashionable. Rust has moved from experiment to infrastructure without becoming a mass-market skill.

Meanwhile, the single biggest lever on pay in the public numbers is not the language but the country and the role: the same back-end job is reported at $175,000 in the US and $79,742 globally. A sensible plan starts from the work you want to do, picks the language that dominates that kind of work, and treats the second language as a hedge. The data above tells you where activity is growing. It cannot tell you who will hire you.

---

Related Articles

• Countries with the Most Software Developers
• Web Development Job Market 2026
• ML Engineer Career Path 2026
• ML Job Market 2026

Sources: Stack Overflow Developer Survey 2025, technology and work; GitHub Octoverse 2025; BLS Occupational Outlook Handbook; Python 3.14 release notes; Linux kernel patch concluding the Rust experiment; Android Kotlin overview; Stanford Digital Economy Lab.

Last verified: October 21, 2026