Worldwide AI Users: Free versus Paid, and Adoption by Country

No company publication read for this report gives paying users by country, and this report does not estimate them. Section 9 explains what was measured and what was not.

Scope: consumer generative AI assistants and chatbots, with the business seats that vendors report alongside them. The data file holds every table in this report, including Microsoft’s full 147-economy table.

1. Summary

These rows use different measures: weekly users, monthly users, paid seats, survey estimates and telemetry models. This report does not add them together or rank them against each other.

Five findings:

• The disclosed paying layer is thin next to the free base. In February 2026, OpenAI reported more than 50 million consumer subscribers and more than 900 million weekly users. That is about 5.6% at that date. OpenAI has reported higher weekly user counts since then, but we found no newer subscriber count, so this report computes no later conversion rate. Of the other assistants covered here, we found a published paying count only for Microsoft 365 Copilot, which is sold as business seats.
• Use is highest in the UAE and Singapore. Microsoft measures the UAE at 73.3% and Singapore at 64.3% of the working-age population using generative AI in Q2 2026. The next five are Ireland, France, Norway, Spain and New Zealand, all between 44% and 50%. The world figure is 18.8%.
• The gap between Microsoft’s Global North and Global South keeps widening. It went from 10.6 points (H2 2025) to 12.1 (Q1 2026) to 12.6 (Q2 2026). Usage was 28.8% in the North and 16.2% in the South.
• Most AI users pay in one surveyed market. Menlo Ventures’ July 2026 survey of US adults finds 55% of AI users pay for at least one AI product. US adult use rose only from 61.0% to 63.6% between its April 2025 and July 2026 survey waves. Menlo reports that "more users became payers". Spending is concentrated: the 14% of payers who spend $100 or more a month account for 60% of consumer AI spending. This is one country’s survey and cannot stand in for others.
• Age divides use sharply in both surveys read here that report it by age. Across six countries in 2025, 47% of people aged 18 to 24 had used ChatGPT in the previous week, against 9% of those 55 and over (Reuters Institute). In the US in February 2026, about two thirds of adults under 30 used AI chatbots, against 23% of those 65 and over (Pew Research Center).

How to read this report. Every figure is dated and sourced to the organisation that published it. Where a number is an estimate or a model, the text says whose and how it was made. Where a sentence is Wavn’s own reading, the text says so. No company publication read for this report gives paying subscribers by country, and this report does not estimate them (section 9).

2. Users by platform

Vendors report different things. OpenAI reports weekly active users. Google reports monthly users of the Gemini app. Microsoft reports paid Microsoft 365 Copilot seats. CNNIC reports an estimate of the number of people in China who use generative AI. The rows below are not comparable with each other.

2.1 ChatGPT, as disclosed by OpenAI (weekly users)

Sensor Tower, an app analytics firm, reports that ChatGPT reached 1 billion monthly active users on its mobile app in May 2026 (press release, 16 June 2026). This is Sensor Tower’s estimate, and it measures something different from OpenAI’s weekly count.

2.2 Gemini app, as disclosed by Google (monthly users)

On these disclosures the Gemini app went from 400 million to more than 1 billion monthly users in about fifteen months. Google’s 11 August post gives one figure by surface: more than 100 million active users on iOS. It does not break down the rest.

2.3 Web and app measurement firms

These are panel and model estimates from measurement firms, not vendor counts.

• Similarweb (published 29 July 2026, updated 17 September 2026). ChatGPT’s share of worldwide generative AI website visits fell from about 76% in June 2025 to roughly 53% in May 2026. Gemini rose from under 9% to around 27 to 28%, and Claude from about 2% to close to 9%. Across June 2025 to May 2026, average monthly web visits to generative AI platforms reached 9.5 billion, unique visitors reached 655 million, and app downloads reached 4.4 billion. These shares cover web visits only.
• Sensor Tower (16 June 2026). ChatGPT’s "True Audience" share, which counts unique users across mobile app and web, fell below 50% for the first time in March 2026.

3. What is disclosed about paying users

OpenAI. On 27 February 2026, OpenAI reported more than 50 million consumer subscribers and more than 9 million paying business users, against more than 900 million weekly active users. OpenAI said January and February 2026 were "on track to be the largest months for new subscribers" in its history. Dividing the two consumer figures gives about 5.6% at that date. Both inputs are "more than" figures, so the ratio is approximate. OpenAI’s 31 March 2026 funding post repeated "over 50 million subscribers". We found no later subscriber count, so no conversion rate is computed against the later weekly figures.

Microsoft. Microsoft reported that Microsoft 365 Copilot "reached over 30 million paid seats" in its fiscal fourth quarter results on 29 July 2026. These are business seats, not consumer subscriptions.

Google, Anthropic and others. We found no published subscriber count for the Gemini app or Google AI plans in the Google posts read for this report, including the 11 August 2026 post announcing one billion monthly users. We found no published user or subscriber count from Anthropic for Claude. We did not search for counts from DeepSeek, Perplexity, xAI, Meta or Chinese domestic services, so this report says nothing about whether they publish one.

3.1 Surveys and market estimates of paying

Menlo Ventures is a venture capital firm that invests in AI companies. Its report marks its portfolio companies, and Anthropic is among them. Its 2026 State of Consumer AI (16 September 2026) is based on a survey of 5,067 US adults conducted with Morning Consult in July 2026. It finds:

• 63.6% of US adults used AI, up from 61.0% in its April 2025 wave.
• 55% of US AI users pay for at least one AI product, which Menlo puts at roughly 90 million Americans. The typical payer spends $20 to $49 a month.
• The 14% of payers who spend $100 or more a month account for 60% of consumer AI spending.
• Payers use AI daily at nearly twice the rate of non-payers (50% against 26%). 65% of payers use an AI agent, against 13% of non-payers.

Menlo also publishes global estimates, built from its US survey and "triangulated against credible third-party sources", with the global figure adjusted down for internet access, age and regional adoption. On that basis it estimates consumer AI spending at $40 billion in 2026, more than three times its $12 billion estimate for 2025, and AI users at 2.0 billion people, up 11% from 1.8 billion. In its 2025 report it estimated that roughly 3% of AI users paid. That 2025 figure is a global estimate, while the 55% is a US survey result, so the two do not measure the same thing.

Sensor Tower said on 16 June 2026 that global in-app purchase revenue from AI apps "is expected to surpass $4 billion in H1 2026", up 36% on the second half of 2025. This was a projection made before the half-year ended. Sensor Tower also estimates that Claude’s average revenue per user on mobile in the US rose from under $0.50 in September 2025 to $2.76 in May 2026.

4. Adoption by country

Four kinds of measure are used below. Microsoft models population use from its own telemetry. The Reuters Institute and KPMG with the University of Melbourne run multi-country surveys. Pew Research Center and CNNIC each measure one country. OpenAI publishes country rankings of ChatGPT messages per person. They measure different things and are not merged.

4.1 Microsoft AI Economy Institute, Global AI Diffusion, Q2 2026

What it measures. Microsoft’s figure is the share of people aged 15 to 64 "who have used a generative AI product during the reported period". It is derived from aggregated, anonymised Microsoft telemetry and adjusted for differences in operating-system and device market share, internet penetration and population. Microsoft calls it "the strongest cross-country measure available today" and also says that "no single metric is perfect". The values are modelled estimates. Microsoft sells AI products of its own, and its data comes from its own software.

Headline figures (published September 2026).

Microsoft reports that almost every economy saw higher usage in the quarter. South Korea had the largest absolute gain (+3.5 points, from 16th to 12th place). Saudi Arabia had the largest rise in rank (30th to 25th). Japan had the largest relative change (+2.2 points, roughly 10% on its Q1 level of 22.5%).

The 30 highest economies, Q2 2026.

Ranks follow the order of Microsoft’s table. Where two economies have the same rounded value, Microsoft’s order is kept.

Every region, Q2 2026. Wavn grouped Microsoft’s 147 economies by UN M49 region, with Western Asia shown on its own. The figures are unweighted: each economy counts once, whatever its population. The full table, with each economy’s region, is in tables.csv.

At the bottom of Microsoft’s table are Tajikistan, Afghanistan and Turkmenistan (each 6.3%) and Cambodia (6.1%).

Reading the table with care.

• Some groups of economies share identical values. Thirteen economies, most of them in West Africa, are at 10.4%. Nine economies, eight of them in Africa, are at 8.0%, and five Central African economies are at 9.1%. The report does not say how these values were derived.
• Microsoft says its current measure leaves out several widely used domestic tools in China. It plans to widen the tool list in its next report and expects measured use to rise in almost every economy, most of all in China. Its 17.5% for China is therefore not comparable with CNNIC’s count (section 4.5).
• Bahrain is not in Microsoft’s table.

4.2 Reuters Institute, six-country survey, 2025

The Reuters Institute’s Generative AI and News Report 2025 (October 2025) is based on a YouGov online survey of about 2,000 people in each of Argentina, Denmark, France, Japan, the UK and the US, fielded from 5 June to 15 July 2025. The sample is non-probability. Across the six countries, the share who had ever used a generative AI system rose from 40% (2024) to 61% (2025). Weekly use nearly doubled from 18% to 34%. ChatGPT was used weekly by 22% on average. Weekly use of AI to get information more than doubled, from 11% to 24%, and overtook creating media (21%).

Use roughly doubled in most of the six countries. The US rose five points, from 31% to 36%, and in 2025 was second only to Argentina.

4.3 KPMG and the University of Melbourne, 47 countries, workplace use

Trust, attitudes and use of artificial intelligence: A global study 2025 surveyed 48,340 people in 47 countries between November 2024 and mid-January 2025. Three in five employees (58%) said they intentionally use AI at work on a regular basis. The share was 80% or more in India, China, Nigeria, the UAE, Saudi Arabia and Egypt, against under 50% in most advanced economies. This is a survey of employees. It is not a population measure and should not be set beside Microsoft’s population figures.

4.4 United States: Pew Research Center, February 2026

Pew surveyed 5,119 US adults from 17 to 23 February 2026, with a margin of sampling error of plus or minus 1.6 points for the full sample. Findings:

• 49% of US adults use an AI chatbot such as ChatGPT, Gemini or Copilot. In August 2024 the figure was 33% on an all-adult basis. The 2024 question was asked only of adults who had heard of chatbots and used different wording, so the two years are not strictly comparable.
• About a quarter (24%) use chatbots daily.
• 44% use ChatGPT, against 34% (2025), 23% (2024) and 18% (2023). Before 2026 the question asked only about ChatGPT and only of people who had heard of it.

4.5 China: CNNIC

The China Internet Network Information Center’s 57th Statistical Report on China’s Internet Development (5 February 2026) reported:

• 602 million generative AI users at December 2025, up 141.7% from the end of 2024, a penetration rate of 42.8%.
• 1.125 billion internet users, a penetration rate of 80.1%.

CNNIC’s Generative AI Application Development Report (2026) was released on 29 September 2026. By CNNIC’s estimate, China had more than 700 million generative AI users by the end of the first half of 2026, with penetration above 50%. Answering questions was the main use: 76.0% of users ask generative AI to answer questions. Processing images or video (47.8%), text (37.6%) and work summaries, minutes or slides (32.5%) followed.

The figure is CNNIC’s estimate. Its definition differs from Microsoft’s (section 4.1) and from vendors’ active-user counts.

4.6 What OpenAI publishes by country

OpenAI’s Signals data (updated 6 August 2026) covers messages on individual ChatGPT plans (Free, Go, Plus and Pro). It ranks countries by ChatGPT messages sent per person and publishes country-level shares of message types. It does not publish users, visits or paying subscribers by country, so it cannot support a traffic or subscriber split by country. What it does report:

• Users who mainly write in a language other than English are now more than half of active users. The leading non-English languages are Spanish, Portuguese and Arabic (30 June 2026).
• Since July 2023, weekly active users have grown fastest, in relative terms, in Africa and Asia, and in lower Human Development Index countries (30 June 2026).
• The ten economies whose per-person message rank rose most from Q4 2025 to Q1 2026 were in Latin America and the Caribbean, Asia-Pacific, Africa and Europe (11 May 2026). OpenAI notes these are changes in rank, not in total usage.

OpenAI’s 6 August 2026 update moved the ranking to Q2 2026: from Q1 to Q2 2026, Peru, Uruguay and Costa Rica rose most, and parts of Latin America, Oceania and Africa rose faster than elsewhere. It did not publish a Q2 table of the kind above.

5. Who uses AI assistants, and for what

5.1 Age

Both surveys read for this report that report use by age, the Reuters Institute’s and Pew’s, find large gaps by age.

• Across the Reuters Institute’s six countries in 2025, 59% of 18 to 24 year olds had used any generative AI in the previous week, against 20% of those 55 and over. For ChatGPT the figures were 47% and 9%. The Institute found smaller age gaps for Gemini and almost none for Copilot and Meta AI, which it links to their being built into widely used products.
• In Pew’s February 2026 US survey, use of AI chatbots was about 66% among adults aged 18 to 29, 61% at 30 to 49, 42% at 50 to 64 and 23% at 65 and over (Pew’s appendix table "Share of Americans who say they use a chatbot, by demographic group").
• OpenAI reports that in Q1 2026 users under 35 still sent the largest share of ChatGPT messages, while users over 35 gained share (11 May 2026). In France and Czechia, the share of messages from users 35 and over rose by more than 10 points in the year to June 2026 (6 August 2026).

Young US adults are also the most sceptical. Roughly half of Pew’s respondents under 30 say AI will have a negative impact on society over the next 20 years, against 14% who say positive. Across age groups, 61% to 65% say AI is advancing too quickly (63% of all adults).

5.2 Gender

Pew finds that the US gender gap in overall chatbot use "has closed": two years earlier, men were 11 points more likely than women to use chatbots. Men remain more likely to use them daily (27% against 20%). Women are about twice as likely to expect a negative rather than a positive impact of AI on them personally (33% against 17%). OpenAI, which infers gender from first names, reports that users with typically feminine names now account for over half of the users whose gender it can infer (Q1 2026).

5.3 What people use it for

In China, CNNIC’s 2026 report finds answering questions the leading use, cited by 76.0% of users (section 4.5). Across the Reuters Institute’s six countries, getting information was the leading weekly use in 2025 (24%). On these three surveys, information-seeking leads, although the surveys ask different questions.

5.4 Time spent and payers

Menlo Ventures’ July 2026 US survey finds that 24% of AI users spend more than an hour a day with AI, 10% more than two hours, and 29% less than 15 minutes. 41% of AI users have tried an AI agent, and 48% say they have earned money with AI.

6. Plans, prices and advertising

Prices below are as published by the vendor on the date shown. Each vendor’s US-dollar price is shown because it is the one price every vendor publishes; local prices are on each vendor’s page. Google says its prices vary by country. OpenAI’s Help Center says ChatGPT purchases are in US dollars, with local-currency billing in a limited set of countries. Anthropic’s page says its prices exclude applicable tax. Prices change: check the vendor’s own page before quoting one.

Perplexity, Grok, Meta AI and the Chinese domestic assistants were not covered in this pricing table.

6.1 Lower-priced tiers

Dates and prices are from OpenAI’s release notes and posts:

• 18 Aug 2025: ChatGPT Go launched in India at ₹399 a month (GST included).
• 22 Sep 2025: Go launched in Indonesia at Rp 75,000 a month.
• 3 Nov 2025: OpenAI began a promotion giving eligible users in India 12 months of Go at no cost. The promotion ended on 21 January 2026.
• 16 Jan 2026: OpenAI said Go had launched in 171 countries since August 2025 and brought it to the US and every country where ChatGPT is available, at $8 a month in the US.

Google launched Google AI Plus in 35 more countries and territories on 27 January 2026, including the US at $7.99 a month. That made it available everywhere Google AI plans are sold. Google says prices vary by country.

6.2 Higher-priced tiers

All three subscription vendors in the table sell tiers above their standard plan. Anthropic’s Max starts at $100 a month. Google added a $100 AI Ultra plan and cut its top Ultra tier from $250 to $200 on 19 May 2026. OpenAI lists Pro plans at $100, $200 and $500 a month (Help Center, read 24 October 2026). Google says its Pro and Ultra subscribers can buy pay-as-you-go credits for Google Antigravity and Google Flow, with the Gemini app "coming soon" (19 May 2026). This blurs the line between subscription and metered use.

6.3 Advertising in ChatGPT

OpenAI is the only vendor whose advertising programme this report covers. Other vendors’ advertising policies were not surveyed. OpenAI’s ads run on the Free and Go tiers only (OpenAI’s posts; its Help Center article on Go says ads may come to Go in future). Plus, Pro, Business, Enterprise and Education have no ads, according to OpenAI’s 9 February 2026 post.

OpenAI describes advertising as "one pillar" of its business, alongside consumer subscriptions, enterprise offerings and usage-based APIs (31 August 2026). In Wavn’s reading, the ads give the free and Go tiers a revenue line of their own, separate from conversion to paid plans.

6.4 Business seats

Microsoft reports more than 30 million paid Microsoft 365 Copilot seats (29 July 2026). OpenAI reports more than 9 million paying business users of ChatGPT (27 February 2026). Consumer subscriptions are therefore one of several revenue lines that vendors disclose, alongside business seats, usage-based APIs and, at OpenAI, advertising.

7. Notes by region

Each note uses the same three sources where they exist: Microsoft’s Q2 2026 diffusion figures, the multi-country surveys, and national surveys. Regions are in alphabetical order.

Africa. Microsoft’s 46 African economies run from South Africa (23.9%) to Rwanda (7.6%), with a median of 10.4%. Many share identical values (section 4.1). Microsoft’s H2 2025 report found DeepSeek, which offers a free chatbot and open model weights, had its strongest adoption in China, Russia, Iran, Cuba, Belarus and across Africa, where it estimated usage at 2 to 4 times other regions. Among employees, KPMG and the University of Melbourne found regular workplace use of 80% or more in Nigeria and Egypt. That is a survey of employees, not a population measure: Microsoft’s population figure for Nigeria is 10.4%. OpenAI reports Africa among the regions with the fastest relative growth in ChatGPT weekly users since July 2023.

Asia (Central, South, East and South-eastern). The range in Microsoft’s table is wide, second only to Western Asia’s: Singapore 64.3%, South Korea 40.6%, Taiwan 33.8%, Vietnam 27.9%, Japan 24.7%, Malaysia 22.6%, the Philippines 20.7%, India 18.5%, China 17.5%, Indonesia 14.8%, Thailand 13.4%, down to Cambodia at 6.1%. South Korea posted the largest gain of any major economy for three reporting periods in a row, and Japan the largest relative gain in Q2 2026. China is measured nationally by CNNIC at more than 700 million users (section 4.5), and Microsoft says its own figure for China undercounts domestic tools. KPMG found workplace use of 80% or more among employees in India and China. OpenAI reports Asia among the regions with the fastest relative growth in ChatGPT weekly users.

Europe. Microsoft’s 32 European economies have a median of 31.0%. The highest are Ireland 49.9%, France 49.6%, Norway 49.4%, Spain 45.1%, the UK 43.7% and the Netherlands 43.5%. Germany (32.3%), Italy (31.4%) and Finland (30.6%) are in the top 30. The lowest is Ukraine at 9.5%. In the Reuters Institute survey, weekly use in 2025 was 34% in Denmark, 28% in the UK and 27% in France. In Pew’s spring 2025 survey across 25 countries worldwide, a median of 53% trusted the EU to regulate AI effectively, against 37% for the US and 27% for China. Pew’s spring 2026 survey of a different set of 12 countries, in Asia, Africa, Latin America and the Middle East, gave medians of 43% for China, 35% for the US and 34% for the EU.

Latin America and the Caribbean. Microsoft’s 24 economies run from Costa Rica (29.7%) to Cuba (6.8%), with a median of 18.7%. Mexico is at 21.4% and Brazil at 20.2%. Argentina had the highest weekly use of the Reuters Institute’s six countries in 2025 (51%), against 23.4% of its working-age population in Microsoft’s Q2 2026 model. The two measure different things: one is a survey of online adults, the other a telemetry model. Five of the ten economies whose ChatGPT per-person message rank rose most in Q1 2026 are in this region: the Dominican Republic, Haiti, Mexico, Brazil and Costa Rica (OpenAI, 11 May 2026); in Q2 2026 Peru, Uruguay and Costa Rica rose most (OpenAI, 6 August 2026).

Northern America. Canada (38.5%, 16th) and the US (33.0%, 21st) are both below most of Western Europe on Microsoft’s measure. Two national surveys cover the US. Pew finds 49% of adults use AI chatbots (February 2026). Menlo Ventures finds 63.6% of adults use AI and 55% of AI users pay for at least one AI product (July 2026). They ask different questions, which is why their adoption figures differ.

Oceania. New Zealand (44.1%) and Australia (40.7%) are in Microsoft’s top 11. Papua New Guinea is at 7.6%, and its per-person ChatGPT message rank rose five places in Q1 2026 (OpenAI, 11 May 2026). OpenAI’s 6 August 2026 update reports parts of Oceania rising faster than elsewhere from Q1 to Q2 2026.

Western Asia. The UAE leads Microsoft’s world table at 73.3%. Qatar is at 43.5%, Israel 38.7%, Saudi Arabia 31.4% (up five places to 25th), Jordan 31.2%, Oman 27.6%, Kuwait 22.6% and Turkey 18.9%. Syria is the lowest at 7.4%. KPMG found workplace use of 80% or more among employees in the UAE and Saudi Arabia.

8. What to watch

• The next diffusion report. Microsoft says its next report will widen the set of AI tools it counts. It expects measured use to rise across almost all economies, most of all in China, and says the change will reflect better measurement rather than a sudden change in adoption.
• A new subscriber count. OpenAI’s latest consumer subscriber figure dates from February 2026, while its weekly user count has risen since. The next subscriber disclosure is needed before any newer conversion rate can be stated.
• Paying beyond one market. The only survey read for this report that measures how many AI users pay covers US adults. Comparable country surveys would allow the free-versus-paid split by country that this report does not attempt.
• Prices by country. Google says AI Plus prices vary by country, and OpenAI launched Go at local prices in India and Indonesia. Neither publishes how many people pay at each price.

9. Method and limits

What was read. For this report, the publishers’ own pages and documents were opened on 24 October 2026:

• company posts, investor releases, help-centre articles and pricing pages from OpenAI, Google, Alphabet, Microsoft and Anthropic
• Microsoft’s Global AI Diffusion reports for Q2 2026 and H2 2025, with the full appendix table
• CNNIC’s pages for its 57th Statistical Report and its Generative AI Application Development Report (2026)
• the Reuters Institute’s 2025 report (PDF)
• Pew Research Center’s Americans and AI 2026 report (PDF and topline), its age chapter, and its global surveys of October 2025 and September 2026
• Menlo Ventures’ 2026 report and methodology
• the KPMG and University of Melbourne 2025 study (PDF)
• Similarweb’s and Sensor Tower’s own publications
• OpenAI’s Signals pages

Section 10 lists every source cited.

Definitions. Weekly active users (OpenAI), monthly users (Google), paid seats (Microsoft), CNNIC’s user estimate, Microsoft’s modelled population share, survey self-reports and measurement-firm panel estimates are different measures. This report never adds them together or ranks one against another.

What was not available, and what was not estimated.

• Paying users by country. No vendor publication read for this report gives them. Estimating them from web-traffic shares was considered and not done: the only country traffic shares found came from a third-party compilation that could not be checked against a primary source, and the weights such a model needs would be assumptions, not measurements. OpenAI’s own country data (Signals) ranks countries by messages per person and does not publish users or subscribers by country, so it cannot replace the traffic shares. This report therefore gives no country split of paying users, of free users, or of ChatGPT traffic.
• Gemini, Claude and most other subscriber counts. We found no published count from Google or Anthropic. Other vendors were not searched.
• US prices for Google AI Pro. Google’s pricing pages served local prices for the location of the reader, and no US price for AI Pro was found in Google’s posts read for this report.
• CNNIC’s full report. CNNIC’s release pages were read. The figures above are from those pages, and the full PDF was not read.
• Region coverage. National surveys of comparable depth were read for the US (Pew, Menlo Ventures) and China (CNNIC) only. Other countries are covered by Microsoft’s model and by multi-country surveys. This reflects what was found and read, not a judgement about which markets matter.

Estimates and modelled figures in this report.

• Microsoft’s diffusion values are modelled from Microsoft telemetry.
• Menlo Ventures’ $40 billion spending, 2.0 billion users and 3% paid in 2025 are Menlo’s estimates.
• Sensor Tower’s in-app purchase figure for H1 2026 is a projection.
• Similarweb’s shares are panel estimates of web visits.
• CNNIC’s 700 million is CNNIC’s estimate.
• Three figures are Wavn’s own arithmetic on published numbers, labelled where they appear: the 5.6% ratio at February 2026, the 850 million weekly users not counted as consumer subscribers, and the regional summary of Microsoft’s table.

Interests. Wavn, Inc. makes AI software and so has a commercial interest in this market. Wavn’s product is built on AI models bought from several of the companies covered here, and its connector works with AI tools from several of them.

This report takes no position on any product, and naming a product here is a statement of a published market fact only. Microsoft and OpenAI publish data about the use of products they sell. Menlo Ventures marks its portfolio companies in its report, and Anthropic is among them.

Written in Wavn from published sources; figures were checked against the sources in section 10 on 24 October 2026.

Sources

1. Scaling AI for everyoneOpenAI · 27 February 2026
2. OpenAI raises $122 billion to accelerate the next phase of AIOpenAI · 31 March 2026
3. Improving GPT-5.6 Sol in ChatGPT, and expanding access to GPT-5.6 Luna for free usersOpenAI · 6 August 2026
4. A milestone in expanding access to AIOpenAI · 31 August 2026
5. DevDay 2026 RecapOpenAI · 29 September 2026
6. ChatGPT Ads expands to 26 new countries as over 100,000 advertisers have used the platformOpenAI · 21 October 2026
7. Testing ads in ChatGPT (updates 26 Mar, 7 May, 11 Aug 2026)OpenAI · 9 February 2026
8. Our approach to advertising and expanding access to ChatGPTOpenAI · 16 January 2026
9. New ways to buy ChatGPT adsOpenAI · 5 May 2026
10. ChatGPT Ads expands across Europe (updated 31 Aug 2026)OpenAI · 18 August 2026
11. ChatGPT Ads expands to Southeast Asia and TaiwanOpenAI · 23 September 2026
12. ChatGPT release notes (entries of 18 Aug 2025, 22 Sep 2025, 3 Nov 2025) (undated page, read 24 October 2026)OpenAI Help Center · 24 October 2026
13. What is ChatGPT Go? (undated page, read 24 October 2026)OpenAI Help Center · 24 October 2026
14. What is ChatGPT Plus? (undated page, read 24 October 2026)OpenAI Help Center · 24 October 2026
15. About ChatGPT Pro tiers (undated page, read 24 October 2026)OpenAI Help Center · 24 October 2026
16. Signals individual dataOpenAI · 6 August 2026
17. How ChatGPT adoption broadened in early 2026OpenAI · 11 May 2026
18. How ChatGPT adoption has expandedOpenAI · 30 June 2026
19. From asking to doing: How the world is putting ChatGPT to workOpenAI · 6 August 2026
20. More than 1 billion people are using the Gemini app every monthGoogle · 11 August 2026
21. I/O 2026: Welcome to the agentic Gemini era (Sundar Pichai remarks)Google · 19 May 2026
22. Alphabet earnings call Q2 2026: Sundar Pichai remarksGoogle · 22 July 2026
23. Google AI Plus is now available everywhere our AI plans are available, including the U.S.Google · 27 January 2026
24. Everything new in our Google AI subscriptions, fresh from I/O 2026Google · 19 May 2026
25. Earnings Release FY26 Q4Microsoft Investor Relations · 29 July 2026
26. Global AI Diffusion: Q2 2026 Trends and Insights (PDF)Microsoft AI Economy Institute · September 2026
27. The continued state of global AI diffusion in 2026Microsoft On the Issues · 21 September 2026
28. Global AI Adoption in 2025: A Widening Digital Divide (H2 2025, PDF)Microsoft AI Economy Institute · January 2026
29. Plans and pricing (undated page, read 24 October 2026)Anthropic · 24 October 2026
30. The 57th Statistical Report on China's Internet Development (release page)China Internet Network Information Center (CNNIC) · 5 February 2026
31. Generative AI Application Development Report (2026) (release page, dated 8 October 2026)China Internet Network Information Center (CNNIC) · 29 September 2026
32. Generative AI and News Report 2025: How people think about AI's role in journalism and society (PDF)Reuters Institute for the Study of Journalism · October 2025
33. Americans and AI 2026: Chatbots, Smart Devices and Views on Impact (report PDF with methodology and topline)Pew Research Center · 17 June 2026
34. How Americans' opinions and use of AI differ by agePew Research Center · 17 June 2026
35. Trust in the EU, U.S. and China to regulate use of AIPew Research Center · 15 October 2025
36. Globally, More People Expect AI to Cause Job Loss Than Growth (PDF)Pew Research Center · 17 September 2026
37. 2026: The State of Consumer AIMenlo Ventures · 16 September 2026
38. Trust, attitudes and use of artificial intelligence: A global study 2025 (PDF)University of Melbourne and KPMG · April 2025
39. AI Search Stats in 2026 (updated 17 Sep 2026)Similarweb · 29 July 2026
40. State of AI 2026 press releaseSensor Tower · 16 June 2026