Is PHP Dead? The 2026 Picture, by the Numbers

I’ve been writing PHP for 18 years. In that time I’ve read the phrase “PHP is dead” at least 50 times — people were saying it in 2008, in 2014, in 2020. The sentence never changes; the evidence never materializes. At some point I stopped getting defensive. Because I realized that arguing without data is just making noise.

This post is going to show you that data. I won’t say “PHP is dead” or “PHP isn’t dead.” I’ll just lay out the numbers. You decide.

What Does “Dead Language” Even Mean? — Defining the Criteria First

Before we get into the argument, let’s define what we’re actually measuring. When is a language considered dead? I see six signals:

1. Usage share is declining — are real-world, running systems dropping the language?
2. Developer popularity is falling — are new developers entering, or are existing ones leaving?
3. Ecosystem activity has stalled — is package/library production continuing?
4. The language itself isn’t evolving — are new versions shipping, and how often?
5. The job market is shrinking — is the number of listings declining?
6. Newcomers aren’t learning it — is a junior pipeline forming?

Let’s measure each one for PHP.

1. Usage Share — W3Techs

According to W3Techs data from May 2026, PHP is used on 71.2% of all websites with a known server-side language. That number has been moving in a narrow band for years — a small drift from the high 70s down to the low 70s — but it’s still four times the usage of the second most common server-side language.

Version distribution looks like this:

PHP Version	Usage
8.x	59.9%
7.x	31.6%
5.x	8.4%
4.x	0.1%

Nearly 60% of sites are on PHP 8.x — modern PHP. 31% are still on 7.x (we’ll come back to that; it’s a warning signal). But the claim that “the ecosystem never moved past 5.x” is completely inconsistent with the data.

For comparison: on the same list, ASP.NET sits at 5.3%, Ruby at 5.1%, Java at 4.2%, Node.js at 3.4%. Even if PHP’s share has declined, the gap between it and its competitors is enormous.

2. Developer Popularity — Stack Overflow & JetBrains

Stack Overflow’s 2025 Developer Survey covered more than 49,000 developers. PHP in the “most-used language” list:

• All respondents: 18.9% (12th place)
• Professional developers: 19.1% (13th place)

That figure is higher than Go (16.4%) and Rust (14.8%) — meaning PHP has more active usage than the two languages developers most often cite as “trendy.” On top of that, JetBrains’ 2025 State of PHP report found that 89% of the 1,720 active PHP developers surveyed are on PHP 8.x, and 88% have 3+ years of experience.

One nuance worth noting: in the Stack Overflow “learning to code” segment, PHP comes in at 14.5%, ranked 15th. So newcomers aren’t gravitating to PHP the way they do to Python or JavaScript — that’s a genuine warning signal for the junior pipeline, but it’s not “dead” territory.

3. Ecosystem Activity — Packagist

Packagist’s live statistics page shows the following numbers (May 2026):

• Total packages: 453,911
• Total versions: 5,578,059
• Total downloads since 2012: 181,729,296,774 (181 billion)

181 billion package downloads means tens of millions of Composer invocations every single day. That figure belongs to an industry-standard ecosystem, not a dead one. For comparison: Perl’s CPAN has ~220,000 modules (half of PHP’s), Ruby’s RubyGems has ~180,000 gems (again, less than half of PHP’s).

4. Is the Language Evolving? — PHP Release Schedule

PHP has shipped a new major version every November for the past six years:

Version	Release Date	EOL
PHP 8.0	November 2020	November 2023
PHP 8.1	November 2021	December 2025
PHP 8.2	December 2022	December 2026
PHP 8.3	November 2023	December 2027
PHP 8.4	November 2024	December 2028
PHP 8.5	November 2025	December 2029

Annual cadence, with concrete new features every release: 8.0 brought JIT, union types, match; 8.1 added enums and readonly; 8.2 added readonly classes and DNF types; 8.3 brought typed class constants; 8.4 added asymmetric visibility and property hooks; 8.5 introduced lazy objects. This is well above the evolution rate considered healthy for a language. (A language people call “dead” typically has a 3–5-year release gap and features that don’t function.)

5. The Job Market

This is the hardest category to measure — there’s no official global dataset. But the observational signals are clear:

• A search for “PHP” or “Laravel” on LinkedIn returns 500+ active listings in Turkey per month
• WordPress (the single engine running on PHP) still powers 43.5% of the web — meaning the sheer maintenance of that ecosystem alone requires a massive workforce
• The freelance/agency market around Laravel and Symfony is growing globally — the JetBrains report shows 12% of PHP developers are independent and 56% work in small teams

I can’t give a precise number, but there’s no concrete data source behind the “job listings are declining” claim either. If anything, maintaining and extending existing systems alone seems enough to keep PHP developer demand alive for years.

6. Are Newcomers Learning It?

From the JetBrains report: developers with less than 6 months of PHP experience went from 2% in 2024 to 4% in 2025 among survey respondents. That’s a doubling. The junior pipeline, at minimum, hasn’t stopped.

On the Stack Overflow side, PHP ranks 15th in the “learning to code” segment — meaning it isn’t a strong entry-level preference. This is a real weakness. New developers typically start with Python or JavaScript; they come to PHP later, usually because the job demands it. The pipeline is slow, but it’s flowing.

---

Summary of the Six Criteria

Signal	Status
Usage share	✅ Leading (71.2% of websites)
Developer popularity	✅ Top 13 (18.9% — ahead of Go and Rust)
Ecosystem activity	✅ 453K packages, 181 billion downloads
Language evolution	✅ Major release every November
Job market	✅ Strong (especially the Laravel + WordPress ecosystem)
Newcomers	⚠️ Slow but flowing (4%, doubled from last year)

Five strong signals, one warning. This is not the picture of a dead language — it’s the picture of a mature, dominant, slightly tired but still-growing language.

What Do Actually Dying Languages Look Like?

Let’s look at two cases for comparison.

ColdFusion

According to Adobe’s EOL schedule: ColdFusion 2021 exited core support in November 2025. ColdFusion 2025 moved to a subscription model — perpetual licensing is gone. There’s an active developer community and a package ecosystem, but:

• It doesn’t appear in TIOBE (not even in the top 50)
• No new frameworks are emerging
• Job listings are mostly about “migrating away from” existing systems

ColdFusion is not technically dead, but you wouldn’t tell a newcomer to learn it. PHP’s picture is very far from this.

Perl and the TIOBE Anomaly

This deserves a brief aside. The most commonly cited source in “PHP is dying” arguments is the TIOBE Index. In May 2026, TIOBE ranks PHP at 18th. On the same list, Perl sits at 11th.

That’s worth noting — because Perl’s TIOBE jump was 17 places in a single year: from 27th to 10th. By TIOBE CEO Paul Jansen’s own explanation, the reason is that Perl sells 4 times more books on Amazon than PHP. Meaning TIOBE isn’t a usage metric — it’s a blend of search results + book counts + course counts. The methodology doesn’t measure real-world usage in the ecosystem.

Anyone who brings TIOBE as proof that “PHP is dead” has simultaneously proven — by the same metric — that Perl is more popular than PHP. To stay logically consistent: either both claims are true, or TIOBE is the wrong tool for this question.

Sources that actually measure real usage, like W3Techs and JetBrains, say PHP is dominant. TIOBE says 18. They’re measuring different things — one counts code-shipping developers, the other counts book-publishing editors.

Let’s Be Honest: PHP’s Real Problems

Up to this point the data has been largely in PHP’s favor. But the goal here is a data-driven analysis. The places where PHP is genuinely weak in 2026:

1. The async story is still second-class. Projects like ReactPHP, Swoole, and FrankenPHP have made async PHP possible — but they’re layers bolted on top of the language, not its natural flow. Writing the same task in Node.js or Go is far more natural. If you’re building a high-concurrency system, PHP is definitely not the first choice.

2. The type system is maturing, but there are no generics. PHP 8.x brought major improvements to type support — typed properties, intersection types, the never type, asymmetric visibility. But generic types still don’t exist. They’re simulated via PHPStan docblocks; there’s nothing in the language’s own syntax. For anyone coming from TypeScript or Rust, this is a significant gap.

3. A large portion of the community is stuck on 7.x. W3Techs shows 31.6% still on PHP 7.x. PHP 7.4 reached end-of-life in November 2022 — meaning millions of websites are running on a version that hasn’t received a security update in four years. This doesn’t mean PHP is dead; it means half of the PHP community is refusing to modernize. That situation keeps the “PHP is an old language” perception alive.

4. It’s the wrong choice for some domains. Real-time messaging, high-frequency trading, machine learning, mobile backends — Go, Java, Rust, or Python are the natural choices there. You can write PHP for those domains, but it’s not where the language shines. CRUD-heavy web applications, content management, e-commerce — there PHP is still comfortably unmatched. We shouldn’t fall into the trap of “every language for everything.”

I can say all of this without hesitation, because my goal here is not to defend a language. It’s to show both strengths and weaknesses based on real data. Whatever side it comes from, an argument made without real data smells like fanaticism.

Closing: Why Is This Such a Sticky Myth?

The data is right there. So why does the “PHP is dead” claim keep cycling back every few years for two decades?

A few hypotheses:

New languages are marketed by “killing” the old ones. When Node.js launched, it was burying PHP. When Go launched, it was burying Python. When Rust launched, it was burying C++. That rhetoric isn’t a technical assessment — it’s market positioning.

The poorly-taught versions of PHP are still lodged in people’s minds. Thousands of developers learned PHP between 2008 and 2014 using PHP 4 and 5.2–5.3 — spaghetti code, global state, security holes were their lived experience. Their mental image of PHP comes from there; the language changed completely after 2015, but the image didn’t.

PHP doesn’t fit the trend-following culture. New framework, new language, new paradigm — marketing those is a profession. PHP sits in the “been doing the same job for years, just doing it better” position. That’s not a compelling story to produce content about; so saying “I write PHP” is not trendy. Not being trendy and being dead are not the same thing.

---

Back to the data: as of May 2026, PHP runs on 71.2% of the web, is used by 19% of developers, its ecosystem has logged 181 billion package downloads, and the language ships a major version every year. At the same time, it’s weak on the async side, has no generics, and half the community is stuck on a four-year-old version.

This is not a “dead” picture. It’s a picture of a mature, dominant, legitimately criticizable language. If you can already see the difference, there was no reason to write this post in the first place. If you couldn’t, at least the sources are here now.

As for me personally: I’ve been writing PHP for 18 years, I’m still writing PHP in 2026, and I’ll most likely still be writing PHP in 2030 — because it’s a good tool that gets the job done. But I also use Go, Python, or Node.js in the same project when they’re the right fit. I don’t do language fanaticism. I read data.

---

Sources

• W3Techs: PHP Usage Statistics, May 2026
• W3Techs: Server-side Programming Languages Overview
• Stack Overflow Developer Survey 2025 — Technology
• JetBrains: The State of PHP 2025
• Packagist Statistics
• PHP Supported Versions
• endoflife.date: PHP
• endoflife.date: ColdFusion
• TIOBE Index
• Slashdot: Is Perl the World’s 10th Most Popular Programming Language?