The rise of the non-English internet

English built the internet. That is a historical fact, not a permanent condition. In 2025, non-English speakers make up more than four out of every five internet users on the planet — and the gap is widening fast. The question isn't whether the internet will become multilingual. It's who builds the infrastructure for when it does.

The numbers

There are approximately 8.1 billion people on Earth. Around 1.5 billion speak English — as a first or functional second language. That's roughly 18% of the global population. Yet English accounts for approximately 54-55% of all content on the web, according to W3Techs data. The math is stark: English has about three times the web presence its speaker population would predict.

The content-to-speaker gap is even more dramatic for specific languages. Hindi is spoken natively by over 600 million people — the third-largest native-speaker community on earth — and accounts for less than 0.1% of web content. That's a 6,000x gap between population and web presence. Arabic: 320 million native speakers, under 1% of web content. Swahili: 200 million speakers across East Africa, a fraction of a percent.

These aren't small populations with niche interests. These are some of the largest language communities on earth — and they are systematically underrepresented on the web that the rest of the world takes for granted.

The languages taking over

Chinese

With over 1.4 billion native speakers and an internet-connected population exceeding 1 billion, Chinese is the largest single-language internet user base by volume. China's domestic internet ecosystem — WeChat, Weibo, Douyin, Baidu — is enormous and largely self-contained, precisely because the global English-first internet doesn't serve Chinese speakers well. When Chinese internet users engage with global platforms, they face a wall. That wall shapes the entire structure of Chinese internet culture.

Spanish

Spanish is the second-largest native-speaker language on earth, with native speakers spread across 20+ countries on two continents. Latin America's internet user base is growing at roughly 7% annually. The Spanish-speaking internet is already substantial — but it's fragmented. A speaker in Mexico, Argentina, and Spain exists in three different cultural contexts using the same language. The opportunity for a truly Spanish-language global social layer is enormous and largely uncaptured.

Arabic

The Arabic-speaking world spans the Middle East and North Africa — a region with young, growing populations and rapidly increasing smartphone penetration. The Arab internet user base is projected to pass 450 million by 2030. Arabic is also right-to-left and has significant dialect variation, making it particularly poorly served by infrastructure built for left-to-right, standardized-form languages. The gap between Arabic-speaking internet populations and Arabic-language web content is one of the widest in the world.

Hindi and South Asian languages

India alone has over 700 million internet users — more than the entire population of Europe. Hindi is the primary language for the largest share of those users, but India's linguistic diversity means that hundreds of millions more are primarily served by Bengali, Tamil, Telugu, Marathi, Gujarati, Kannada, and dozens of other languages. The Indian subcontinent is the single largest underserved language market on the internet. Major platforms have attempted Hindi localization with mixed results; the smaller language communities remain almost entirely unaddressed.

Portuguese

Brazil is the largest Portuguese-speaking country and one of the world's largest internet markets — over 180 million users. Brazilian internet culture is vibrant, mobile-first, and growing. But Portuguese-language content on major platforms is dwarfed by English. Brazilian creators who want global audiences must produce in English. Brazilian businesses entering global markets face the same localization overhead as any other non-English company.

Swahili and African languages

Africa deserves its own paragraph because the scale of what's coming is not yet reflected in current data. The continent has 1.4 billion people, a median age of 18, and an internet user base growing at roughly 10% per year — three times the global average. By 2030, Africa is projected to have the world's largest population of first-time internet users. The majority of those users will not be English speakers. They will be Swahili, Hausa, Amharic, Yoruba, Zulu, Igbo, and dozens of other language communities. The infrastructure that serves them — or fails to — will be built in the next five to ten years.

Why most apps still default to English

The honest answer is: because that's how they were built, and the cost of change is high.

The internet was bootstrapped by a small number of engineers, mostly in California, who spoke English. The tools they built — databases, frameworks, CMS platforms, ad networks, SEO infrastructure, developer documentation — were built in English. The default assumptions baked into every layer of web infrastructure are English-language assumptions: left-to-right text, 26-character alphabets, keyword search optimized for English morphology, ad targeting built on English-language content signals.

As these tools scaled into global platforms, the English default scaled with them. Localization was added as an afterthought — a translation layer bolted onto a monolingual foundation. The result is that even platforms with hundreds of millions of non-English users still make product decisions, run company communications, and structure developer communities in English first.

"We localized the interface. We never localized the culture."
— A common observation from engineering leads at global platforms attempting to expand beyond English-speaking markets

There is also a measurement problem. Engagement metrics, revenue data, and growth signals are most clearly visible in English-speaking markets, which are also the most heavily instrumented. Non-English markets are often measured less precisely, creating a feedback loop where investment follows legible data, which is disproportionately English-language data. The non-English internet is systematically underinvested in because it's systematically undermeasured.

The cost of that default for businesses

The business case for multilingual product strategy has never been stronger — and the cost of ignoring it has never been higher.

Research on consumer behavior consistently shows that people are 3 to 4 times more likely to make a purchase when they can do so in their native language, and significantly more likely to return to a service that communicates with them in their language. For e-commerce, the data is unambiguous: localized product pages, customer support, and checkout flows drive materially better conversion and retention than English-with-translation-bolted-on.

Beyond direct conversion effects, there's a market access issue. Companies that do not localize effectively are structurally restricted to the fraction of any non-English market that is comfortable enough in English to complete a purchase, file a support ticket, or engage with product documentation. For most emerging markets, that fraction is a small minority. The majority of the addressable market is behind a language wall the company built itself.

These are not marginal users. These are hundreds of millions of people with smartphones, internet connections, and purchasing power — who are systematically excluded from most global digital services because those services were built in English and never fully left it.

What this means for product builders

The shift is not hypothetical and it is not slow. The non-English internet is the internet, now. Product builders who internalize this ahead of the curve will have structurally better access to the majority of internet growth over the next decade. Those who treat English as the default and everything else as optional will find themselves managing an increasingly small slice of a much larger market.

The practical implications are significant:

• Onboarding and UX must work in native languages from day one — not as a premium or a later phase, but as a baseline expectation for any product targeting global users.
• Support infrastructure must be multilingual — users who cannot get help in their language churn faster and at higher rates.
• Content and community must be language-native — a Spanish-language community grafted onto an English-language platform is not the same as a community built for Spanish speakers from the start.
• Analytics and measurement must account for non-English signal — if you can't measure behavior in your non-English user populations, you can't optimize for them, and you will systematically underinvest in them.
• Real-time communication infrastructure must be language-agnostic — the expectation that users in a product will share a language is becoming less valid with every passing year.

The last point is the hardest to address with existing infrastructure. Chat, communities, social features, multiplayer experiences — all of these historically assumed shared language. Real-time multilingual communication at scale is not a feature. It's a new infrastructure layer.

How Babel is built for this shift

Babel is built on the premise that the non-English internet is the primary internet, not the edge case. Every architectural decision in Babel starts from the assumption that users will communicate in different languages simultaneously — and that the job of the platform is to make that invisible.

When a user posts in Hindi, a follower in Brazil reads it in Portuguese. When a community in Nigeria creates a group, members from Japan and France can participate — each in their own language. When someone sends a voice message in Arabic, the recipient hears it in Spanish. The translation is not a feature the user invokes. It's infrastructure that runs beneath every interaction, automatically.

This isn't just a translation layer bolted onto a social network. It's a fundamentally different architecture: one where language is a per-user attribute of the communication, not a property of the content. Content exists once; it's rendered in each user's language independently. A community built on Babel is not constrained by the language distribution of its members in a way that communities on any existing platform are.

The non-English internet is the next chapter of the web. Babel is the infrastructure for it.