Ask ten people which country is winning the AI race and you’ll probably get ten confident answers. The truth shifts depending on what you count: money, models, patents, or the number of ordinary people who actually open a chatbot on a given week. In 2026 those measures no longer point in the same direction.
This list looks at five places that lead on at least one meaningful measure, then eight where usage data suggests people are still waiting for AI to arrive. The gaps are measured mostly through Stanford’s 2026 AI Index and Microsoft’s AI Diffusion reports. Every number comes with its source, and the limits of each measure are noted where they matter.
Leader 1: United States, Where the Money and Machines Are
Stanford’s 2026 AI Index reports that US private AI investment reached $285.9 billion in 2025, roughly 23 times China’s $12.4 billion, and the US hosts 5,427 data centers, more than 10 times any other country.[1] Oxford Insights, which tracks how ready governments are for AI, also ranked the US first again in its latest Government AI Readiness Index.[2] By almost any measure of capital and infrastructure, nothing else comes close.
Model output is strong too, though counts differ by method. The report’s figures put US notable model releases in 2025 at roughly 50 to 60, depending on how Epoch AI defines a notable model.[3] The country’s weak spot is everyday use. The US dropped to 24th place globally in population level adoption, at 28.3%, despite leading in AI investment and model development.[4]
Leader 2: China, Closing the Gap Without the Cash
The headline finding in Stanford’s 2026 report is how little separates the top American and Chinese models. The performance gap has collapsed to 2.7%, down from 17.5 to 31.6 percentage points in May 2023.[5] DeepSeek-R1 matched the top US model in February 2025.[1]
China also leads on several research measures. It accounts for 69.7% of global AI patent filings and 23.2% of publications, and installs industrial robots at nine times the US rate.[5] The private investment gap overstates things somewhat, because government guidance funds are estimated to have put roughly $184 billion into AI firms between 2000 and 2023, outside the private investment tally.[3] Adoption among the public is another matter, since China sits at a relatively low 16%.[6]
Leader 3: South Korea, Small Country, Dense Innovation
South Korea doesn’t dominate headlines the way Washington or Beijing does, yet the data keeps flagging it. Stanford found that South Korea leads the world in AI patents per capita.[7] The same report counted eight notable models from Korea in 2025, behind the US with 59 and China with 35.[8]
Ordinary use is climbing quickly as well. Microsoft’s data shows South Korea’s AI diffusion rising from about 26 percent in the first half of 2025 to about 31 percent in the second half,[9] the largest jump among the economies in its fast risers table. Another analysis notes that South Korea is now the second largest ChatGPT subscriber market globally, behind only the United States.[10]
Leader 4: United Arab Emirates, Adoption by Design
When the question is how many people actually use generative AI, the UAE tops the chart. Microsoft’s report shows adoption rising from 59.4 percent to 64 percent between the first and second halves of 2025,[9] or nearly two thirds of the population. A later 2026 mapping of the same type of data put the UAE above 70% adoption, with Singapore next at 63%.[6]
This isn’t an accident of wealth alone. Microsoft’s report gives the country its own section on what it calls a deliberate AI strategy. The UAE also leads a table of 147 economies, which Quartz describes as led by the UAE, Singapore and Ireland in the share of working age people using generative AI.[11]
Leader 5: Singapore, Proof That Size Isn’t Destiny
Singapore is tiny next to the superpowers, but its usage numbers rival anyone’s. Rest of World, summarizing Microsoft’s research, notes that over 60% of the population in both the UAE and Singapore used some type of generative AI tool in the past year.[12] One breakdown puts the figure at 60.9 percent.[13]
The 311 Institute’s Global AI Power Index, which ranks 93 countries, argues that Singapore shows a smaller country can compete through strong government, investment and real world use.[14] That’s one index with its own weighting, so treat it as a useful lens rather than a final verdict. Still, a city state getting this far on usage alone says something about how deliberate policy can shape adoption.
Behind 1: Japan, Rich, Advanced, and Still Cautious
Japan is the most surprising name on the laggard side. Microsoft’s data shows adoption at 19.1% in the second half of 2025, up from 16.7%,[15] and the latest quarterly report puts it at 24.7% in the second quarter of 2026,[16] roughly one quarter of the population. That’s low for one of the world’s wealthiest economies.
The picture isn’t fully bleak. Microsoft says Japan’s global diffusion ranking improved from 56th in the first half of 2025 to 48th in the first quarter of 2026,[17] helped by AI models that handle Japanese much better. Measurement matters too: a self reported Japanese government survey of people aged 15 to 69 found 58.8%,[18] a far higher number from a different method.
Behind 2: Mexico, Climbing From a Low Base
Mexico’s usage has been rising, but slowly. Microsoft measured 17.8% in the second half of 2025, up from 16.7%,[15] and 21.4% by the second quarter of 2026.[16] That means a little over one in five people in the working age population now use generative AI.
Regional context helps explain the pattern. Rest of World reported that in South and Central America, 14 of 24 countries had adoption rates below 20%, including Mexico, Venezuela, and Brazil.[12] The same coverage says most adoption in the region is driven by government agencies, while private companies lag because of high costs and a lack of technical skills.[12]
Behind 3: Brazil, Big Market, Modest Usage
Brazil is Latin America‘s largest economy, yet its AI use has trailed the leaders by a wide margin. Microsoft’s data moves it from 15.6% in the first half of 2025 to 17.1% in the second half,[15] and then to 20.2% in the second quarter of 2026.[16] About one fifth of working age Brazilians now use these tools.
Connectivity isn’t the whole story. World Bank data gives Brazil an internet usage rate of 84.2% in 2023,[18] which suggests access alone doesn’t explain the gap. Another factor is where the models come from: the US and China together account for 74% of AI models, leaving other economies far behind.[6]
Behind 4: Philippines, Gaining Slowly While Policy Moves
The Philippines shows how adoption and policy can move at different speeds. Usage rose from 17.1% to 18.3% across the second half of 2025,[15] then reached 20.7% in the second quarter of 2026,[16] about one fifth of the population. Its quarterly gain of 0.6 points was the smallest among the five economies in that Microsoft table.
On the government side, the country is moving up. Oxford Insights notes that the Philippines rose seven places in the latest readiness index and will chair ASEAN, possibly shaping an ASEAN approach to AI.[19] Rest of World adds that India, the Philippines and Malaysia saw slight improvements in AI usage, with ChatGPT and Gemini among the most widely used tools.[12]
Behind 5: Nigeria, Africa’s Biggest Population, Thin Usage
Nigeria has the largest population in Africa, but Microsoft’s numbers put usage near one in ten. The figure was 9.3% in the second half of 2025,[15] then 10.4% in the second quarter of 2026.[16] One independent review cautions that the Nigeria value is itself a West Africa average shared by twelve countries.[18]
Access is a big part of the story. The same review cites World Bank data showing internet usage of 40.1% in Nigeria in 2023.[18] Policy is moving regardless, and Oxford Insights points to the launch of Nigeria’s AI Scaling Hub[19] as one example of strategy turning into practice.
Behind 6: Kenya, Barely Moving in the Latest Data
Kenya’s figure in Microsoft’s latest quarterly tables sits below one in ten people. It was 8.7% in the first quarter of 2026 and 8.9% in the second,[16] a rise of only 0.2 points. That’s about the smallest change among the economies in the lowest group.
A caution applies here. An analysis of Microsoft’s methodology says thirty six rows in the data are regional averages shared with neighboring countries,[18] and several African economies in the table carry identical numbers. Kenya’s figure should be read as a rough regional signal. The wider divide is clearer: Microsoft found that adoption in the Global North grew almost twice as fast as in the Global South, widening the gap from 9.8 to 10.6 percentage points.[15]
Behind 7: Bangladesh, Half the Global Average
Bangladesh sits near the bottom of Microsoft’s second quarter 2026 table. Usage rose from 7.8% in the first quarter to 8.0% in the second.[16] For comparison, global adoption reached 16.3% of the working age population in the second half of 2025.[20]
That puts Bangladesh at roughly half the global average, though the two figures cover different periods. Microsoft’s technical paper explains that South Asia and Sub Saharan Africa remain below 13% in AI adoption.[21] The researchers also note that limited AI presence in low income countries makes quantitative study difficult because national data is scarce.[21]
Behind 8: Ethiopia, Among the Lowest Readings Anywhere
Ethiopia closes the list with one of the lowest usage rates Microsoft tracks. In the second half of 2025, a summary of the report placed it among the lowest adopters at 6.8%, alongside Cambodia, Afghanistan and Rwanda.[20] By the second quarter of 2026, its figure had risen from 7.6% to 8.0%.[16]
Roughly one person in twelve now uses generative AI there, according to that estimate. Microsoft’s own conclusion on the trend is that AI’s benefits are expanding, but they are not expanding equally.[15] Ethiopia’s numbers show that gap in its plainest form.
