The Multilingual Internet—Ram Mohan on Digital Inclusion and Global Growth at CloudFest 2026

Language determines inclusion. This is true in the real world, and even more so online. When information appears in your native language, you trust it more and engage with it on a deeper level.

The challenge today is that the internet is consolidating around a few dominant languages. Historically, there were perhaps seven major languages online. With the rise of AI and large language models, we are now seeing a convergence toward two primary foundations: English and Chinese.

Even when AI responds in another language, it often processes the request in English before translating the result. This creates subtle barriers and biases. If your native language is far from these dominant linguistic structures, accessing knowledge becomes harder. We must move toward a world where language no longer hinders participation in digital life.

Consider it from a business perspective. You might deploy an AI agent that works perfectly in English, German, or Spanish. But when you expand into markets in Africa or Southeast Asia, your tools suddenly fail to reach large portions of the population.

Today, digital expansion requires significant customization, but it should not. When systems genuinely understand local languages and cultural contexts, businesses can enter new markets with far greater ease. Building language capability into the design stage is not just a choice for inclusivity; it is a choice for business value and growth.

Security itself is a language problem. Take phishing. In English, we understand the metaphor immediately. But how do you explain “phishing” in a language where the word only relates to catching fish?

If users lack the vocabulary to understand digital threats, they cannot protect themselves. Governments delivering online services face the same issue. Security warnings, authentication terms, and compliance language created mostly in English do not always translate cleanly. We have to design systems and terminology that make sense globally, not just assume translation will solve the problem.

Absolutely. A great example is universal acceptance. Years ago, when I ran the .info registry, many websites rejected email addresses ending in .info because their systems assumed valid domains had only two or three letters. We had a situation where technically valid identities were being excluded.

We worked across the industry to fix that, eventually getting major companies to update their systems to recognize modern and multilingual domain names. But here is the interesting part: the phrase “universal acceptance” itself does not translate cleanly into many languages. The concept exists, but expressing it requires cultural adaptation. That experience reinforced a fundamental truth: language cannot be an afterthought. It must be embedded at the architectural level from Day One.

The goal of good infrastructure is invisibility. It should enable change without causing disruption. When we migrated millions of domain names, the real success was that nobody noticed. Websites stayed online, and businesses continued operating normally.

Today, language inclusion is becoming another layer of infrastructure. If we get it right, billions of users will gain access without even realizing the complexity behind the scenes.

Expect a conversation about where the internet needs to go next. We will explore where the next wave of meaningful growth will come from. The biggest untapped opportunity is not a new platform or protocol; it is the billions of people who remain underserved because language was treated as a feature, not as foundational infrastructure.

If we are serious about building the next generation of digital services, especially in the AI era, then language cannot sit at the edge of the system. It must be built into its core.