ChatGPT Statistics (2026): Usage, Trend, Market & Growth

The most reliable ChatGPT statistic in this article is not a user count. It is $40 billion — OpenAI's annualised revenue run rate as of August 2026, confirmed by CNBC and Bloomberg ahead of a planned IPO. Financial figures reported to investors carry legal exposure if misstated, which makes them the strongest tier of evidence available here.

User numbers are a different matter. You will see 900 million, one billion and 1.1 billion quoted interchangeably, often in the same article, because they measure different things over different windows. This piece keeps those separate and labels every figure by how much weight it deserves.

The most interesting finding, though, is neither of those. It is that enterprise revenue overtook consumer revenue for the first time in July 2026 — which means the most famous consumer product in AI now makes most of its money from businesses.

The Headline Numbers

User Growth: Roughly Doubling Year on Year

Weekly active users went from 400 million in February 2025 to 900 million in February 2026 — more than double in twelve months — and reached approximately one billion by July 2026.

Two caveats. Only three reliable datapoints exist across eighteen months, so the line between them is interpolation rather than measurement — the real curve may not be smooth. And the growth rate is clearly decelerating in percentage terms: doubling in the first twelve months, then roughly 11% over the following five. That is what saturation looks like at a billion users, and it is normal rather than alarming.

Revenue: Flat, Then a Step Change

The revenue picture is more dramatic than the user picture. Run rate sat around $24–25 billion from March through the spring of 2026, then accelerated sharply — reportedly growing 20% month over month in July — to reach $40 billion by August.

Run rate is not annual revenue. It annualises a recent month or quarter, so in a fast-growing business it runs well ahead of what was actually collected over the preceding twelve months. A $40 billion run rate does not mean $40 billion was earned in the last year.

The flat spring followed by a sharp summer step is the interesting shape. Something changed in the middle of 2026 — and the enterprise data below suggests what.

The Enterprise Flip

In July 2026, enterprise revenue surpassed consumer revenue for the first time, with business customer count growing 32% in a single month and enterprise settling above 50% of the mix.

That reframes what ChatGPT actually is commercially. It remains the most widely used consumer AI product on earth by a wide margin — around a billion weekly users — but the revenue now comes predominantly from businesses buying seats, not from individuals buying Plus.

The arithmetic is worth sitting with. Business customers are roughly 18% of paying accounts but produce over half the revenue, which tells you the average business seat is worth several times the average consumer subscription. It also explains the revenue step change without any corresponding jump in user growth.

The Numbers That Get Confused

Most ChatGPT statistics articles quote user figures that are not comparable, then treat them as if they are. These are four different populations measured four different ways:

Why this matters when you read any AI usage statistic:

• Monthly and weekly actives are not the same population. The gap between 1.1 billion monthly and roughly 1 billion weekly implies a very high return rate, which is genuinely impressive — but quoting one as the other inflates or deflates by around 100 million people.
• Actives are not customers. Roughly 50 million paying users against a billion weekly actives is a conversion rate of about 5%. That is respectable for a freemium consumer product and it is a very different number from "a billion users".
• Third-party estimates are the weakest tier. The 1.1 billion MAU and 1.44 billion download figures come from analytics panels rather than from OpenAI, and panel methodology varies.
• Dates matter more than usual here. A 900 million figure and a 1 billion figure are both correct — five months apart. Articles that omit the date make them look contradictory.

Market Share

ChatGPT is estimated to hold roughly 77% of the generative AI market by third-party measurement. Treat that as directionally true and methodologically loose: what counts as "the generative AI market" varies enormously between measurement firms, and a figure that high usually reflects consumer web traffic rather than enterprise API spend.

For contrast, enterprise API spend tells a different story — that market is far more contested, and our Claude AI statistics piece covers the survey data showing how differently it splits. Both things can be true: overwhelming consumer share, contested enterprise share.

What the Numbers Actually Say

1. The revenue acceleration is real and well-sourced. $24 billion in March to $40 billion in August, reported to investors ahead of an IPO. This is the most trustworthy figure in the set.
2. User growth is decelerating, and that is expected. Doubling to 900 million, then 11% over five months. You cannot keep doubling at a billion.
3. The business has quietly become an enterprise business. July 2026 is the inflection: enterprise above 50% of revenue while representing under a fifth of paying accounts.
4. Conversion is around 5%. Fifty million payers against a billion weekly actives — the free tier is doing enormous distribution work.
5. Run rate is not revenue. Every headline number here annualises a recent period rather than reporting a completed year.

How to Read AI Statistics Without Getting Burned

Pro tip: before citing any AI usage figure, check three things — who measured it, exactly what they counted, and what date it refers to. Most apparent contradictions in AI statistics are two correct numbers from different months.

The tiering that applies across this whole category: company-confirmed financials are strongest, because misstating them to investors carries legal consequences. Statements from company spokespeople are next. Reported figures attributed to named outlets with investor sourcing follow. Third-party panel estimates are weaker. Aggregator numbers with no traceable origin are weakest of all, and they are unfortunately the most widely republished.

If a statistic is load-bearing for a decision you are making, trace it to a primary source before you rely on it.

Keep Reading

More data and pricing: Claude AI stats 2026, ChatGPT pricing explained, Claude pricing, Grok pricing, what ChatGPT Work is, and how to use ChatGPT for content creation. Or browse all guides and prompts on PromptsRush.