◆TDW

Remote Developer Salaries by Stack: What the 2024–2026 Data Shows

Median pay by language and infrastructure tool from three years of Stack Overflow survey data, how companies actually price remote engineers, the Europe–US gap, and why your stack moves your pay far less than where you live and how senior you are.

"Which language pays the most?" is one of the most searched questions in software careers, and most answers recycle a single chart. The data from 2024 to 2026 gives a less exciting but more useful answer: the language you write moves your pay by tens of percent, while your location and seniority move it by multiples. This article lays out the numbers by stack, explains why the niche languages top the tables, and covers what that means for remote engineers, especially in Europe.

Where the numbers come from

Stack Overflow's annual Developer Survey is the largest public source of salary by technology. Since 2025 it no longer publishes a "top paying technologies" table as text, so every per-language figure from 2025 onwards that you see quoted is someone's recalculation from the raw data. The figures below were calculated from Stack Overflow's public response files for 2024, 2025 and 2026. Each is the median self-reported total annual compensation in US dollars (salary plus bonuses, before tax) of everyone who said they worked with that language in the past year. The calculation reproduces Stack Overflow's own published 2024 figures exactly (Erlang $100,636, Elixir $96,000). The salary samples are 23,435 respondents in 2024, 23,947 in 2025 and 13,507 in 2026.

Pay by language

Language2024 global2025 global2026 globalUS 2025Remote 2025
Ruby$90,221$102,961$110,000$170,000$124,437
Scala$88,619$96,524$91,132$200,000$115,216
Elixir$96,000$99,194$90,982$175,000$110,214
Go$76,433$89,815$92,000$175,000$113,085
Rust$76,292$83,000$84,666$160,000$114,496
TypeScript$65,907$79,648$85,096$155,000$92,576
Python$67,723$76,828$83,022$150,000$96,146
C#$66,066$78,616$83,644$140,000$94,618
JavaScript$63,694$75,410$80,000$150,000$90,000
Java$61,714$75,109$76,513$150,500$92,812
PHP$49,586$60,000$64,458$124,400$69,676
Dart$44,576$50,000$54,294$152,000$60,328
All respondents$65,000$75,320$79,610$150,000$88,492

Median total compensation, USD. "Remote" means respondents who work fully remote; hybrid is excluded.

Three patterns hold every year. The mainstream languages — JavaScript, TypeScript, Python, Java, C# — sit within about 10% of the overall median. A handful of smaller communities run 25–40% above it: Erlang, Elixir, Scala, Ruby and Clojure (Erlang and Clojure are left out of the table because their samples are small; the 2025 survey did not offer Clojure as an answer at all). PHP and Dart are at the bottom in every year.

Inside a single country the spread shrinks. Among US respondents in 2025, JavaScript, TypeScript, Python, Java, Kotlin and Swift all fall between roughly $150,000 and $162,500. The real outliers are Scala at $200,000 at the top and C# ($140,000) and PHP ($124,400) below the US median. In Germany the 2025 medians for TypeScript, JavaScript, Python, Java and Rust are identical, at $81,210, which matches a round €70,000 at the survey's exchange rate: within one market, many developers on different stacks report the same salary.

Infrastructure tools pay more consistently than languages

The steadiest premium in the data is not a language. Terraform is the highest-paid widely used technology in all three years: $93,099 (2024), $104,413 (2025) and $110,316 (2026) globally, and about $185,000 among US respondents. That is 30–40% above the global median. Kubernetes ($92,812) and AWS ($92,000) follow in 2025. Other sources point the same way: Dice's 2025 US report (means, not medians) puts Docker at $132,051 and DevOps at $131,947 against an overall average of $112,521, and Arc.dev's 2026 freelance data puts DevOps and platform work at $90–170 an hour against $55–120 for frontend.

Why the niche languages top the table

The high medians for Erlang, Elixir, Scala and Ruby mostly describe who uses those languages, not what learning them pays. In the 2025 data, Erlang users had a median of 14 years of experience and Ruby users 13, against 11 for all respondents. In 2024, 60–65% of Erlang and Elixir users worked fully remote against 41% overall, while PHP and Dart users were the least likely to be in the US and Dart users were the most junior, at a median of six years. Small samples also make niche rankings unstable: Erlang has moved between about $99,000 and $120,000 from one survey to the next.

Treat circulating "top language" lists with care. Several 2024 and 2025 articles name Zig the best-paid language. That is the 2023 figure (about $103,600). In 2024 Zig was mid-table at $75,332 and in 2025 at $81,210.

Other surveys measure different things and partly disagree. JetBrains reports the share of each language's users who are among the top earners rather than a salary: Scala leads at 38% in 2025, followed in 2024 by Go and Kotlin. Stack Overflow agrees on Scala but puts Go and Kotlin mid-table globally. Dice reports a C# premium in the US, while C# sits below the US median in every year of Stack Overflow data. No source controls for country, seniority and role together, so no published figure isolates the effect of the language itself.

Remote pays more in surveys, but not for the same job

In the raw 2025 data, fully remote respondents report a median of $88,492 against $46,406 for in-office respondents, and the gap appears within every language. Most of it is composition: remote workers are more senior and far more likely to be in high-pay countries. In the 2025 survey, 45% of US developers were fully remote against 22.5% in Germany. Within the US the gap narrows to about $15,000 ($165,000 for remote against $150,000 for all respondents).

The best controlled evidence comes from Cullen, Pakzad-Hurson and Perez-Truglia, who studied offers made to 1,396 US tech workers through levels.fyi (NBER working paper, published in AEA Papers and Proceedings, 2025). Comparing offers at the same company, title, level and location, remote positions paid about 1.1% more than office ones. The candidates, meanwhile, accepted about 25.6% less pay to work remotely. The "remote engineers earn 22% more" figures that circulate online come from uncontrolled comparisons.

How companies price remote engineers

Most employers set remote pay by location. In Pave's 2025 data, 88% of companies managing pay bands adjust for location; Carta reports 84–85% and an ERI survey 81%. Payscale's 2026 report puts the share paying a single headquarters or national rate at 26%. These come from different samples, so the honest range for "no location adjustment" is somewhere around 12–26%.

The common models are:

  • Location tiers. In the US, Tier 1 cities (San Francisco, New York, Seattle) at 100%, Tier 2 at 90–95%, Tier 3 at 80–85% (Pave).
  • Nearest office. Pay is set by the closest office, typically within about 50 miles.
  • A national remote rate, usually 85–90% of the highest-cost location.
  • One rate everywhere. 37signals says it pays "at the top 10% for our industry at San Francisco salary levels, regardless of where you live." Buffer sits in between with two bands, Global at 90% and High at 100%.

GitLab, long the reference example of formula-based location pay, retired its public compensation calculator on 30 May 2025 in favour of market ranges per role and location. At companies with location bands, moving somewhere cheaper usually means a pay cut; Google's tool showed cuts of up to 25% in 2021.

Europe and the US

Every source puts European software pay at roughly half of US pay, with the exact ratio depending on what is measured:

  • levels.fyi (crowdsourced, total compensation including equity, skewed towards large tech companies): Europe-wide median $86,811 from 22,239 submissions, against a US median of $190,000–195,000 in 2025, so about 45%. By country: Switzerland $155,379, Denmark $115,018, Ireland $95,830, UK $91,749, Netherlands $88,560.
  • Stack Overflow 2025 (self-reported total compensation): back-end developers at $175,000 in the US, $108,913 in the UK, $87,011 in Germany and $71,929 in France.
  • Pave (employer data, base salary for the same job and level): non-US individual contributors at 40–67% of US pay.

Seniority is the bigger lever. levels.fyi's 2025 US medians run from Entry $155,000 to Mid $226,000, Senior $312,000, Staff $457,000 and Principal $551,000: Senior is double Entry and Staff is nearly triple. Specialisation sits between seniority and language. In Stack Overflow's 2025 US data, AI/ML engineers report $189,500 against $175,000 for back-end and $138,000 for full-stack, but the European AI premium shrinks to under 10% in Berlin data that compares people at the same level.

Ranked by size, the evidence puts location first (about 2x), seniority second (2–3x from Entry to Senior or Staff), specialisation third (0–35%) and language last.

What you'll see in job listings

Whether you can see pay before applying depends heavily on where the employer is. On Indeed, 59% of US postings showed pay in May 2025, and 63% in software development. In March 2026 the share was 56% in the UK, 48% in the Netherlands, 43% in France, 39% in Ireland, 36% in Italy, 17% in Spain and 12% in Germany. Software development is among the least transparent categories in Indeed's European data.

The law is moving at different speeds. In the US, California, Colorado, Hawaii, Illinois, Maryland, Massachusetts, Minnesota, New Jersey, New York, Vermont, Washington and DC require ranges in job postings, with Virginia added from 1 July 2026, and most of these laws also cover remote roles that could be done in the state. The EU Pay Transparency Directive (2023/970) requires employers to give the starting pay or range before the first interview and bans questions about pay history, but it does not require the salary in the ad itself. Its transposition deadline was 7 June 2026; by late September 2026 only Greece, Italy, Lithuania, Malta and Slovakia had fully transposed it.

Using this when you compare offers

  • Ask how the company sets remote pay. Location tier, national rate or one global rate makes a bigger difference than any stack premium. For engineers in Europe, the real opportunity is landing on a higher-paying employer's band, not remote work as such.
  • Compare the same measure. Total compensation with equity (levels.fyi) is not comparable with the base-salary ranges in job ads.
  • Check the level an offer is pitched at. A level is worth more than a language.
  • Know your arrangement. Contractor rates and employee salaries are not comparable either; see contractor, EOR or direct employee.

On The Dream Work, every listing must state a salary range: companies cannot publish a role without a minimum and maximum. The job match score compares the top of each role's range with the minimum you set in your profile, so roles that can't meet your number score lower before you spend time applying. You can browse tech jobs with the salary shown, check whether you're qualified before you apply, or start with the guide on where and how to find remote tech jobs.

Limits of the data

Stack Overflow's respondents are self-selected, recruited mainly through Stack Overflow's own channels, and only 48.8% of 2025 respondents answered the pay question. Pay is self-reported total compensation, converted to dollars at a single exchange rate on one date (25 June 2025 for the 2025 survey) with no adjustment for living costs, so non-US medians move with the exchange rate as well as with the job market. One respondent counts towards every language they list. The 2026 survey is about half the size of 2025. levels.fyi is crowdsourced and skews towards large tech companies. No source found separates the effect of a language from country, seniority and role, so every ranking above shows who uses each stack rather than what learning it would earn you.

Sources

For candidates

See your match score on every open role.

Build your profile in 5 minutes and get scored against every active role.

Create free account →