Every listicle I have read this year about web programming languages picks a winner. Most of them are wrong, but they are wrong in an interesting way. The datasets that journalists cite when declaring JavaScript the champion or Python the ascendant language are measuring different things, and the contradictions are not bugs. They are the actual story.
There is the question of what runs the web that already exists, and the question of what working developers actually touch. Those are two different questions. The first belongs to the people maintaining and fixing what is already deployed. The second belongs to the people picking tools for the next project. Pretending one answer covers both is the mistake most hot takes make.
Where the installed-base number comes from
W3Techs does daily crawls of public websites and tries to identify what server-side language is behind each one. Late September 2026 readings put PHP at roughly 70% of detectable sites. JavaScript sits at single digits. Ruby just lost second place to JavaScript earlier in the year. Python is barely a blip in this dataset.
The honest caveat is that W3Techs only counts sites it can fingerprint, so the headline number is among detectable sites rather than every site on the internet. Even with that caveat, the trend line has been stable for years. The installed web runs on PHP because the installed web is mostly WordPress, and WordPress runs on PHP.
What this number tells you: if you are doing maintenance, security work, plugin development, or rescue jobs on real websites, the language is almost always PHP. If you are picking a stack for a brand new project with no legacy baggage, you are free to ignore the installed base entirely.
Where the developer-survey number comes from
The 2025 Stack Overflow Developer Survey shows the opposite distribution. JavaScript led at 66% of respondents who did extensive work with it. Python jumped year over year. PHP barely registered. The methodology is a self-selected pool of about 32,000 developers who answered a survey about what they touched in the past year.
What this number tells you: when developers pick tools for greenfield work, they pick JavaScript and Python. Almost no developer reaching for PHP on a new project unless the project is explicitly a WordPress site.
Both numbers are right. Each one is right for a different audience.
WordPress is the actionable story
The most useful number in this whole mess is that WordPress powers around 40% of all websites W3Techs can identify. That is the floor of “who has problems you can solve for money.” Most business owners who run WordPress sites cannot write CSS, cannot patch plugins, and do not know what a PHP version upgrade means until the site breaks.
If you can keep a WordPress site hardened, backed up, and online, you have a permanent customer base. The reputation for low rates comes from hobbyists who compete on price. Specialists who compete on outcomes, security, performance, and reliability charge real money for the same tool. The pitch is not “I build WordPress sites.” The pitch is “I will harden your site, set up automatic backups, lock down admin access, and keep it from getting hacked again.” Same tool, very different billable rate.
This is one of the few areas in 2026 web work where the supply of competent help is smaller than the demand. The reason is that most working developers would rather learn something trendier. The gap is your opportunity.
Freelance is real but the math is honest
About one in seven developers in the Stack Overflow survey identify as freelance or self-employed. That is a significant population, not a fringe lifestyle. The honest math, though, is that freelance web coding rates on Upwork cluster around the high $20s to low $30s per hour. Specialists in security, performance, headless WordPress, or AI integration bill multiples of that.
The cold-start problem is real. Freelancers without portfolios, case studies, or warm referrals grind at the bottom of the rate distribution. Specialists with proof bill two to three times the median. Expect to spend a year building the proof before the rates move.
Self-employment costs are the part nobody quotes you. Health insurance, retirement contributions, quarterly taxes, software subscriptions, and unpaid time add up to roughly 30% of gross income for a US freelancer. The headline rate of $30 an hour translates to around $33,000 take-home at 30 billable hours a week, less than a junior salaried developer at most US metros. Freelancing is a tradeoff of autonomy and ceiling upside for the loss of employer benefits and stability.
The salaried side of the data
The May 2025 BLS wage data for web developers and digital interface designers puts the median annual wage in the high $80,000s. The top 10% of earners clear six figures. BLS projects roughly 5% job growth from 2025 to 2035, which is about average for the broader economy. Translation: web coding is not a boom field, not a dying field, and not a field that rewards coasting on credentials.
The salary ceiling in the field goes to people with one of these bet stacks:
- TypeScript and modern JavaScript frameworks for application work
- Python for data, AI, and automation-adjacent roles
- Infrastructure skills like cloud, containers, and CI/CD pipelines
- Security engineering layered on top of any of the above
Pair any of those with a specialty in security or performance, and your compensation moves meaningfully above the median.
Trade-offs
Three costs to be honest about:
Time. The early-career path that combines JavaScript plus a framework plus Python plus a specialty is a multi-year arc, not a one-year sprint. Anyone promising you a job-ready web developer in three months is selling you something. The credential alone does not pay.
Money. Freelance income at the median rate, after self-employment costs, lands around $33,000 take-home for a US freelancer billing 30 hours a week. The ceiling is real but the floor is real too. Decide whether the tradeoff makes sense before you quit a salaried job.
Specialization risk. The boring specialty that pays your bills today may not pay your bills in ten years. WordPress is unlikely to vanish, but it is not guaranteed to remain 40% of the web. The narrow specialty you pick locks in some career risk. Diversify slowly as you build reputation.
Anyone optimizing for income this year should freelance WordPress work. Anyone optimizing for a 20-year career should learn JavaScript and Python deeply and pick a specialty that travels between stacks. Anyone hiring should expect candidates to be JavaScript-heavy and PHP-light, and adjust job posts accordingly.
What I would tell past me
Three lessons, addressed to the version of me that read “language of the year” headlines and made career moves accordingly:
- Bookmark W3Techs and the Stack Overflow survey directly. Every time someone tells you a language is dying or winning, check the actual data before believing them. The data is free and the question is small.
- Calculate the freelance math for your own situation. Include taxes, insurance, retirement, software, and unpaid time. Numbers that look attractive at a headline rate often fall apart after self-employment costs.
- Pick a specialty and write down three problems you could solve in it. That list becomes your next pitch, your next conversation, or your next job interview answer. Specialists earn more than generalists in every survey I have read.