US and Chinese companies train almost all of the world’s most-used AI models | Our World in Data
Subscribe<br>Donate
Data InsightsUS and Chinese companies train almost all of the world’s most-used AI models<br>June 30, 2026<br>US and Chinese companies train almost all of the world’s most-used AI models
Edouard Mathieu<br>Download
Dozens of companies worldwide develop large AI models, but it can be difficult to get a sense of where the most-used ones tend to come from.<br>OpenRouter is a large platform that allows users to interact with and write software on top of AI models through a single interface. It includes all the models from large companies like Google, OpenAI, Anthropic, Meta, DeepSeek, and Alibaba, as well as many, many others.<br>I analyzed the data published by OpenRouter on the 50 most used models each day since January 2025 and calculated the average monthly presence by origin country of the models.<br>As you can see on the chart, US-based companies still account for most models in OpenRouter’s top 50. But their presence has declined, and China-based companies have grown rapidly, from 5 models in the daily top 50 at the beginning of 2025 to 20 in May 2026.<br>Very few top-50 models come from companies outside the United States and China. Canada was represented early in 2025 by Cohere’s Command R models, while France remains represented by Mistral AI’s NeMo model.<br>A technology that more people use every year is, so far, almost entirely the product of two countries.<br>Read more Data Insights on innovation and technological change
Related topic pages:<br>Artificial Intelligence
Copy link
Our latest Data Insights<br>See all Data Insights<br>June 27<br>Obesity rates in Pakistan have tripled in the last 20 years<br>Hannah Ritchie<br>In the 20th century, obesity was a health problem that mostly affected today’s rich countries. In the 21st century, it has emerged as a health challenge almost everywhere. Many low- and middle-income countries have seen rapid increases in the share of people defined as overweight or obese.<br>In the chart, you can see this rise for Pakistan. In the last 20 years, the share of adults defined as overweight has more than doubled, and the share defined as obese has more than tripled. Now, almost 60% of adults are overweight, and one-quarter are obese.<br>These are modeled estimates from the World Health Organization based on national surveys of height and weight. Body Mass Index (BMI) is a person’s weight in kilograms divided by the square of their height in meters. An adult is defined as overweight if their BMI is 25 or greater, and obese at 30 or greater.<br>Pakistan’s own domestic surveys often report even higher figures; that’s because they have a lower BMI threshold for being overweight and obese, which reflects the fact that South Asian populations can have higher health risks at a lower level of BMI.<br>Obesity is a health problem because it increases the risk of developing a range of conditions, such as cardiovascular disease, strokes, diabetes, and various cancers.<br>Explore how obesity rates have changed in your own country
Continue reading
June 25<br>Europeans consume more milk and dairy products than people in other regions<br>Hannah Ritchie and Pablo Arriagada<br>In Europe, the supply of dairy products is far higher than in other parts of the world.<br>This chart shows milk supply per person across the world’s regions, with data from the Food and Agriculture Organization of the United Nations.<br>It’s based on supply — the amount of milk available for consumption — which includes consumer waste, so it is slightly higher than the amount that people actually consume.<br>As you can see, the per capita supply in Europe is more than two and a half times the global average. People across Africa and Asia have far less available.<br>These numbers include dairy products such as yogurt and cheese, based on how much milk is used to produce them, but they do not include butter.<br>Explore country-by-country data on milk supply
Continue reading
June 23<br>A woman’s risk of dying in pregnancy or childbirth varies hugely by country<br>Esteban Ortiz-Ospina<br>How likely is it that a 15-year-old girl will eventually die from a pregnancy-related cause?<br>Researchers at the UN and the World Bank combined available birth and mortality data with statistical models to answer this question. Their estimates assume that the country’s fertility and mortality rates remain constant throughout the teenager’s lifetime (an important assumption I’ll get to later). The chart shows their results.<br>In Chad, the Central African Republic, and Nigeria, the estimated lifetime risk is around 4%. This is dire: it means about 1 in 25 girls would eventually die from a pregnancy-related cause.<br>Women in many other African countries also face substantial risks, and much of Sub-Saharan Africa has a rate above 1%. By comparison, estimates in most other regions are considerably lower, and across Europe the risk is below 0.1%.<br>The very high risks for the countries on the left of the chart...