Translation Statistics 2026: The Two Charts That Decide Your Language Strategy

How big is the language industry now?

Translation has moved from useful to load-bearing. The global language services market is approaching $75 billion, roughly £59 billion, and growing at around five to six per cent a year. Recession, a pandemic and economic turbulence have all failed to slow it, because globalisation and digital trade keep generating material that has to work in more than one language.

Chart one: the languages the web is written in

Here is where the famous figure misleads. You will often read that English makes up around half of all websites. That count assigns a single language to each site, which flatters English and ignores the many sites that publish in several languages at once. On a measure built to handle multilingual sites, the picture is far closer. English sits at about 20 per cent of web content, with Chinese a step behind at roughly 19. Spanish is a clear third near eight per cent. Then comes a long tail, Hindi, Russian, Arabic, French and Portuguese, each between three and four per cent. English leads, but it no longer rules.

Chart two: the languages people actually use

On the same modelling, the audience lines up a little differently. Chinese moves to the front of the connected population at around 18 per cent, with English just behind at roughly 15. Spanish again holds third near seven per cent, then Arabic, Hindi, Russian, French and Portuguese, clustered between three and four. The order barely changes from the content chart, and that is the first thing worth noticing. For the largest languages, supply and demand have mostly drawn level.

Two facts still deserve a place in any planning meeting. English is the mother tongue of only about one person in twenty worldwide, and its share of internet users has been sliding, not climbing, as hundreds of millions come online across Asia, Africa and Latin America.

The gap is the opportunity

So where is the opportunity, if content now tracks the audience for most major languages? In the places where it does not. Two patterns stand out. First, some large audiences remain under-supplied. Arabic and Hindi both claim a bigger share of users than of content, and across much of Africa and Southeast Asia the gap is wider again, as connectivity races ahead of the material people can actually read. Second, English is over-produced, holding more of the world's content than of its users, which means a page in English competes against far more rivals than a page in a language whose readership has outgrown its supply. The lesson is not to translate into everything. It is to back the languages where your buyers already are and the content around them is still scarce.

What is pulling demand upward

A handful of forces keep the pressure on.

• E-commerce. Most online shoppers prefer to buy in their own language, and many will leave a site that does not offer it. Multilingual product pages have stopped being a nicety.
• Streaming and media. The big platforms spend heavily on subtitling and dubbing, and audiences now expect content to travel across borders by default.
• Regulated work. Legal, medical and life sciences translation remains one of the largest and most resilient segments, because in those fields an approximation is a liability.
• Technology. AI and machine translation have lifted throughput sharply. Used well, they let teams handle far more volume, with skilled linguists steering tone, accuracy and cultural fit. The tool carries the volume. People carry the judgement.

Where the growth is

Europe and North America still account for most global spend. The fastest movement is in Asia Pacific, led by China, India and Southeast Asia, where rapid digitisation is pulling demand in both directions, into local languages and back out into English and other major markets.

What this means for your language roadmap

Three conclusions to act on.

• English alone leaves money on the table. More than four in five internet users are not reading in English, and that share keeps growing.
• Quality decides whether the spend pays back. Raw machine output can cost you the trust it was meant to build. Copy that reads as though it was written for the market, not aimed at it, is what converts.
• Pick languages where demand outweighs supply. Match a large online audience with comparatively little existing content, and you reach people while the field is still open.

The bottom line

Both charts tell a calmer truth than the headlines. The web is far less English-bound than the usual statistics suggest, and for most major languages content and audience now sit close together. The advantage goes to whoever spots the places they still part company, the large readerships whose content has not caught up, and to whoever translates well enough to be chosen once they arrive. Match your language choices to genuine audience demand, prioritise the markets where good content is scarce, and invest in translation and localisation that reads as though it belongs.

If you want help deciding which languages to translate first, or website translation built to perform in the markets you are targeting, Brightlines can help. Get in touch for a quote and a clear plan.

Charts and sources

• Web content and internet users by language: OBDILCI model, version 6, July 2025. Shares are modelled on a first-plus-second-language basis, with a confidence interval of plus or minus 20 per cent
• Market size: Mordor Intelligence, Language Services Market, 2026

Methodology note: both charts draw on the same source and method, so they can be read side by side. OBDILCI measures content and users while allowing for the fact that many sites publish in more than one language, which avoids the over-counting of English found in measures that assign a single language to each website.