OpenAI data shows ChatGPT's fastest growth is now non-English

The News

More than half of the people actively using ChatGPT for personal tasks now do so predominantly in a language other than English. That is the headline finding from a new adoption study published by OpenAI, which tracked how its consumer audience has changed since July 2023.

The report draws on a sample of roughly 0.1% of accounts and covers only individual plans (Free, Go, Plus and Pro), excluding Codex, Enterprise and education products. After six months on the service, a typical user sends around 50% more messages each day and attempts twice as many distinct tasks, a sign that early curiosity is hardening into daily habit.

Geography is where the shift is starkest. OpenAI reports that the fastest relative growth is now coming from Africa and Asia, and from countries lower on the Human Development Index rather than the wealthy, English-speaking markets that seeded the product. Spanish, Portuguese and Arabic are the most common non-English languages, while Uzbek, Kazakh and Burmese posted the sharpest increases among languages with more than a million users as of June 2026. The company also notes that people with typically feminine names now lead usage, having reached parity last year.

Why It Matters

Taken together, the numbers describe a product that has quietly outgrown its Silicon Valley origins. For most of the modern software era, consumer platforms were built English-first and localised later, treating other languages as an afterthought bolted on once the American market was saturated. OpenAI's data suggests the opposite is now happening: the growth engine sits in the global South, and the English-speaking core is no longer where the story is being written.

There is a useful precedent here. When mobile internet adoption exploded across emerging economies in the 2010s, it did not mirror the desktop web that came before it. Users leapfrogged straight to cheap smartphones and data, and the platforms that won, from messaging apps to payments, were the ones that designed for that reality rather than porting a Western product wholesale. The same pattern is now visible in generative AI, and it rewards companies that treat non-English users as the centre of the market rather than its periphery.

Indian Angle

For India, this is close to a validation of a bet several local firms have already placed. India is among ChatGPT's largest single audiences, and it sits squarely inside the Asian growth surge OpenAI is describing. The finding that most usage is now non-English gives commercial weight to the wave of Indic-language model work under way at start-ups such as Sarvam AI and Ola's Krutrim, both of which have argued that Hindi, Tamil, Bengali and other Indian languages are underserved by English-centric systems.

It also sharpens the pricing question. OpenAI's lower-cost Go tier, offered at rupee-friendly rates, looks less like a goodwill gesture and more like a direct response to exactly the low-income, high-growth markets this report highlights. For policymakers, the data feeds squarely into the IndiaAI Mission and MeitY's Bhashini programme, which are pouring public money into Indian-language datasets and models. If adoption is genuinely being driven by non-English speakers, the state's language-first strategy starts to look commercially prescient rather than merely nationalistic.

FAQ

What did OpenAI's report actually measure?

It analysed how the consumer ChatGPT audience has evolved since July 2023, using a small sample of individual accounts. It looked at language of use, message volume, task variety and the geography of growth. It deliberately excluded business, developer and education products, so the picture is of personal, everyday usage rather than corporate deployment.

Does it break out Indian-language usage?

Not by name in the top-line figures, which highlight Spanish, Portuguese and Arabic. But India falls inside the Asian region OpenAI singles out for the fastest relative growth, and the broader non-English trend directly implicates Hindi and other major Indian languages.

What does it mean for Indian AI start-ups?

It strengthens the commercial case for Indic-language models from firms like Sarvam and Krutrim, and for consumer pricing pitched at value-conscious users. Demand for vernacular AI is no longer a thesis; it is showing up in a global incumbent's own usage data.

Where can I read the original report?

OpenAI has published the full adoption study on its official site, linked in the attribution below.

This story was reported by OpenAI. Read the full original coverage at OpenAI.