Claude AI Statistics (2026)

A Claude figure can look definitive right up to the moment you ask what it counts. A web visit is not a person, an app MAU estimate is not an Anthropic disclosure, and a revenue run rate is not audited annual revenue. Treating them as interchangeable is how a single number becomes more convincing than useful.

The short answer: Claude AI statistics in 2026 show rapid commercial growth, substantial measured web and mobile reach, and a usage mix that extends beyond coding. But Anthropic has not published an official global Claude MAU total in the 2026 disclosures and research cited below. If you need a citable Claude number, use the measure that matches your question rather than repeating an unsupported total.

Key Claude AI statistics for 2026

• Anthropic reported an annualized revenue run rate of about $9 billion at the end of 2025 and $14 billion in February 2026. Both are company-reported run-rate figures, not audited annual revenue. Anthropic’s February 12, 2026 Series G announcement also said Claude Code’s run-rate revenue exceeded $2.5 billion.

• On April 6, 2026, Anthropic reported that its annualized revenue run rate had passed $30 billion and that more than 1,000 customers were spending over $1 million annually. This customer count covers high-spend customers, not total accounts or Claude users. Anthropic’s Google and Broadcom announcement is the primary disclosure.

• Anthropic announced $65 billion in Series H funding at a $965 billion post-money valuation on May 28, 2026, and said its run-rate revenue had crossed $47 billion earlier that month. Funding and post-money valuation are financing measures, not revenue or a public-market capitalization. Anthropic’s Series H announcement is the source.

• In Anthropic’s sample of roughly one million Claude.ai and first-party API conversations from February 5–12, 2026, Computer and Mathematical occupations accounted for 35% of Claude.ai conversations. This is a share of sampled conversations, not a share of people using Claude. Anthropic’s March 2026 Economic Index documents the sample and analysis.

• Similarweb estimated that claude.ai averaged 331 million monthly visits and 37 million monthly unique visitors worldwide from June 2025 through May 2026. These are third-party web measurements, not a total Claude audience. Similarweb’s September 2026 comparison also estimated Claude app MAU rose from 7.3 million in June 2025 to 103.9 million in May 2026.

• Sensor Tower reported that Claude’s U.S. mobile average revenue per user rose from less than $0.50 in September 2025 to $2.76 in May 2026. This is a mobile monetization measurement, not Anthropic-wide revenue or revenue per Claude user. Sensor Tower’s State of AI 2026 release provides the estimate.

• Anthropic’s June 2026 Claude Code study analyzed about 400,000 interactive sessions from about 235,000 people between October 2025 and April 2026. It reported average Claude Code use of 20 hours per week within that study population, not among all Claude Code users. Anthropic’s Claude Code research explains the dataset and findings.

• GCN reported a Menlo Ventures estimate that Anthropic held 40% of enterprise large-language-model API spend in 2025. The estimate combines a survey of nearly 500 U.S. enterprise decision-makers with bottom-up modeling; it measures enterprise API/platform spend, not consumer chatbot share. GCN’s September 2026 report also notes that Menlo Ventures is an Anthropic investor.

These figures describe different populations and systems. They are useful precisely because they are not forced into a made-up total.

Get the Mobile Testing Playbook Used by 800+ QA Teams

Discover 50+ battle-tested strategies to catch critical bugs before production and ship 5-star apps faster.

Email addressGet Instant Access

100% Free. No spam. Unsubscribe anytime.

What Anthropic’s commercial disclosures measure

Anthropic’s 2026 disclosures show an unusually fast commercial ramp. The sequence—about $9 billion at the end of 2025, $14 billion in February, more than $30 billion in April, and more than $47 billion in May—is a sequence of annualized revenue run rates.

A run rate annualizes revenue from a recent period. It is a directional measure of current commercial pace, not a statement that Anthropic earned that amount during a completed fiscal year. You should use it when discussing momentum, but not when comparing audited annual revenue across companies.

The high-spend customer figures offer a separate enterprise signal. Anthropic said that more than 500 customers spent over $1 million annually in February 2026; by April, it said the number exceeded 1,000. Neither number tells you how many businesses have an Anthropic account, how many employees use Claude at work, or how many consumers use Claude.

The Series H terms are separate again. A $65 billion financing round and a $965 billion post-money valuation show the price investors agreed to in that financing event. They should not be presented as Claude revenue, user growth, or market capitalization.

Why the distinction matters

A compact claim such as “Claude is worth $965 billion and has $47 billion in revenue” changes the meaning of both measures. The valuation is post-money financing value; the $47 billion figure is an annualized run rate reported earlier in May 2026. Keeping those labels in place makes the figures usable in a board deck or market comparison.

How people used Claude in Anthropic’s research

Anthropic’s Economic Index research addresses behavior rather than the size of Claude’s audience. Its March report examined a February 5–12, 2026 sample of approximately one million conversations from Claude.ai and first-party API traffic. A conversation sample can reveal what people ask Claude to do; it cannot establish global MAU.

Coding remained a major use case. Computer and Mathematical occupations represented 35% of Claude.ai conversations in the February sample. Yet Anthropic’s research also points to a broader task mix: the top 10 O*NET tasks made up 19% of Claude.ai conversations in February 2026, down from 24% in November 2025.

That pattern is consistent with Anthropic’s earlier January Economic Index, which found that mostly coding-related Computer and Mathematical tasks fell from 40% of observed tasks in March 2025 to 34% in November 2025. Anthropic’s January 2026 report is a study of observed task patterns, not a census of Claude users.

Anthropic also modeled task coverage rather than adoption. It reported that about 49% of jobs had at least one-quarter of their tasks performed using Claude in its analysis. That does not mean 49% of workers use Claude. It means the model identified Claude use across at least one-quarter of tasks associated with those jobs.

Geographic measures need the same care. In the February analysis, the top 20 countries accounted for 48% of per-capita Claude usage. This is a concentration measure in Anthropic’s analysis, not a count of people in those countries.

Anthropic’s June Economic Index added two useful behavioral views. It reported interviews with 81,000 Claude users in December 2025, a research sample rather than a total user base. It also found personal use in sampled chats and Cowork conversations rose from around 35% on weekdays to just under 50% on weekends. The June 2026 Economic Index supports both findings.

Claude Code statistics show longer workflows

Claude Code has both a commercial disclosure and a separately defined usage study. Anthropic said Claude Code’s run-rate revenue exceeded $2.5 billion in February 2026 and that its weekly active users had doubled since January 1. The company did not disclose an absolute weekly-active-user count, so the percentage change cannot be converted into one.

The October 2025–April 2026 research sample provides more detail on behavior. Anthropic reported that roughly 56% of the 400,000 studied Claude Code sessions involved writing, fixing, or testing/orchestrating code. Within the session mix, fixing declined from 33% in October 2025 to 19% in April 2026, while operating software grew from 14% to 21%.

Those shifts suggest that the observed workflows were moving beyond isolated code generation. They do not prove that every Claude Code user runs agentic workflows, but they do show that operating software accounted for a growing portion of sessions in Anthropic’s studied population.

Anthropic also cited a recent analysis estimating that Claude Code authored 4% of public GitHub commits worldwide. The analysis is unnamed in Anthropic’s report, so this is an attributed estimate—not an independently verifiable Anthropic measurement—and should be cited that way.

Claude’s web traffic, app MAU, and mobile monetization are different measures

Similarweb’s figures provide a view of Claude’s consumer-facing surfaces. Its 331 million monthly visits and 37 million monthly unique visitors cover claude.ai worldwide over June 2025 through May 2026. Visits can include repeat activity; unique web visitors do not cover every mobile-app, API, or embedded-product interaction.

Its 103.9 million May 2026 app-MAU estimate is another measurement layer. It covers the Claude mobile app as measured by Similarweb, and some people may also be counted among the web audience. Adding 37 million web unique visitors to 103.9 million app MAU would manufacture a total from overlapping, differently measured populations.

Sensor Tower’s U.S. mobile ARPU estimate answers a third question: monetization per mobile user. The increase from less than $0.50 in September 2025 to $2.76 in May 2026 may indicate stronger mobile monetization in that market. It does not tell you Anthropic’s global ARPU, its API revenue per customer, or the number of people using Claude.

What the enterprise-spend estimate says—and does not say

The 40% enterprise figure is often the closest available statistic to an enterprise market-share claim, but its scope is narrow. GCN reported that Menlo Ventures estimated Anthropic’s 2025 share of enterprise LLM API spend at 40%, up from 24% in 2024 and 12% in 2023.

The estimate is based on a survey of nearly 500 U.S. enterprise decision-makers and bottom-up market modeling. It measures API and platform spend in enterprises; it does not measure chatbot traffic, free-plan usage, consumer preference, or the overall AI market.

You should also preserve the conflict disclosure: Menlo Ventures is an Anthropic investor. That does not automatically invalidate the estimate, but it is material context when you cite a modeled market-share figure.

Why there is no single verified Claude user total

The widely repeated “245 million Claude users” figure is not a safe substitute for an official user disclosure. The secondary pages that repeat it do not establish it as an Anthropic-published global MAU number, and the primary Anthropic disclosures cited in this report do not publish one.

The disagreement is not necessarily evidence that one provider is wrong. It often means the providers counted different things:

1. Web visits count activity on claude.ai, including repeat visits.

2. Web unique visitors estimate distinct visitors seen on the web surface.

3. App MAU estimates monthly activity in the mobile app.

4. Conversation samples analyze a defined subset of conversations.

5. Enterprise API spend estimates customer expenditure, not usage by individual people.

A person can appear in more than one of these systems. You may visit claude.ai, use the Claude app, and access Claude through a workplace API workflow. Without a published deduplication method across surfaces, adding the metrics together would create a false precision.

That limitation is also the practical answer to the query behind most Claude AI statistics searches: use official commercial disclosures for Anthropic’s business scale, usage studies for behavior, and third-party panels for the specific web or app surface they observe. Do not label any of them as a confirmed global Claude MAU total.

Methodology and source notes

This report prioritizes Anthropic’s own disclosures for commercial figures and usage research. Similarweb and Sensor Tower are measurement providers, so their web, app, and ARPU figures remain labeled as third-party measurements. The enterprise-spend number is a Menlo estimate reported by GCN and is labeled accordingly.

The definitions matter:

• Annualized revenue run rate: revenue pace projected from a recent period, rather than audited revenue for a completed year.

• Monthly visit: a web-traffic event or session, not necessarily a distinct person.

• Monthly unique visitor: a measurement-provider estimate of distinct web visitors, not all product users.

• App MAU: a measurement-provider estimate of monthly active mobile-app users.

• ARPU: average revenue per user for the defined product surface and geography.

• Conversation sample: analyzed conversations from a defined time and source, not a count of all users.

• Modeled task coverage: an analysis of tasks that can be performed using Claude, not a workforce adoption rate.

• Enterprise API/platform spend: enterprise expenditure on the relevant AI access and platform category, not consumer chatbot share.

No first-party Quash data covers Claude adoption, revenue, or usage. The useful differentiator here is therefore not a proprietary total; it is keeping each public statistic attached to the population and measurement method it actually describes.

Conclusion

Claude’s 2026 statistics point to a business growing quickly in enterprise revenue while reaching a large, separately measured web and mobile audience. Anthropic’s research also shows a strong developer center of gravity alongside broader work and personal-use patterns.

For your own citation, start with the decision you are trying to support. Cite a run rate for commercial momentum, a conversation sample for task behavior, app MAU for mobile reach, or API-spend estimates for enterprise position. If your claim requires a global Claude MAU, the defensible answer remains that Anthropic has not publicly disclosed a verified total.