Claude & Gemini Statistics 2026: Revenue, Users and Reach

ChatGPT gets the headlines, but the two assistants behind it are where the most interesting numbers now sit. Anthropic's Claude has become the default model of the enterprise software stack, and Google's Gemini has become the most widely distributed AI assistant on Earth by riding Android, Search and Workspace.

The pace is hard to overstate. Anthropic went from roughly $1 billion in run-rate revenue at the start of 2025 to $14 billion by February 2026, and Google says Gemini's user base tripled in a year.

This article pulls together the verified numbers on both, from Anthropic's own funding disclosures and Economic Index research to Alphabet's earnings commentary, and looks at what the two trajectories tell us about where AI usage is actually heading.

Key Claude and Gemini statistics for 2026

• Anthropic reached $14 billion in run-rate revenue by February 2026, having grown more than 10x annually in each of the past three years, according to Anthropic.
• Anthropic raised a $30 billion Series G at a $380 billion post-money valuation in February 2026, per Anthropic's own announcement.
• Anthropic told investors it expects roughly $10.9 billion in revenue in Q2 2026 alone, and its first operating profit, according to TechCrunch.
• Claude Code passed $2.5 billion in run-rate revenue by February 2026, and Claude models authored around 4% of all public GitHub commits worldwide, per Anthropic.
• Anthropic's customers spending over $1 million a year grew from 12 to more than 500 in two years, per Anthropic.
• Anthropic held 32% of enterprise LLM API usage in mid-2025, ahead of OpenAI on 25% and Google on 20%, and 42% of the code generation market, per Menlo Ventures.
• Coding is 36% of all Claude.ai usage, and 77% of business API usage shows automation patterns, per the Anthropic Economic Index.
• More than 120,000 enterprises now use Gemini, and Gemini Enterprise paid seats grew from 2 million to more than 8 million in a single quarter, per Alphabet.
• Google processed over 1.3 quadrillion tokens a month across its surfaces by Q3 2025, more than 20x annual growth, per Alphabet.
• The Gemini app was downloaded more than 137 million times on iOS in the 12 months to July 2026, with users tripling year on year, per TechCrunch.
• At Gemini 3's launch in November 2025 the app had 650 million monthly users and AI Overviews had 2 billion monthly users, per Google.
• Google began rolling Gemini out to roughly 4 million General Motors vehicles in April 2026, per TechCrunch.

How much money does Anthropic make in 2026?

Anthropic hit $14 billion in annualised run-rate revenue by February 2026, up from roughly $1 billion at the start of 2025 and $5 billion in August 2025. It expects around $10.9 billion of revenue in Q2 2026 alone, alongside its first operating profit.

When Anthropic announced its Series G in February 2026, it disclosed a $14 billion run rate and said revenue had grown more than 10x annually in each of the past three years [1]. The earlier Series F announcement put run rate at "approximately $1 billion" at the start of 2025 and "over $5 billion" by August 2025 [2].

By May 2026, TechCrunch reported figures Anthropic had shared with investors: roughly $10.9 billion projected for Q2 2026, more than double the previous quarter, with the company's first operating profit expected [5].

Two caveats. These are run-rate figures and investor projections, not audited annual revenue, and run rate flatters fast-growing companies by annualising their best recent month. The profit claim comes with a warning too, since TechCrunch noted Anthropic may not stay profitable through the year given its scheduled compute costs [5].

Even discounted, though, this is one of the fastest revenue ramps in software history. The comparison worth making is that a company most consumers have barely used now books more quarterly revenue than many FTSE 100 firms make in a year.

What is Anthropic worth, and who is funding it?

Anthropic raised a $30 billion Series G at a $380 billion post-money valuation in February 2026. By April 2026, Google had committed up to $40 billion more, and reporting suggested investor appetite at valuations of $800 billion or higher.

The Series G, announced by Anthropic on 12 February 2026, valued the company at $380 billion post-money [1]. Barely two months later, TechCrunch reported Google investing $10 billion immediately with up to $30 billion more tied to performance targets, alongside 5 gigawatts of TPU compute over five years [6].

The same report said investors were courting Anthropic at valuations of $800 billion or more, and that an IPO could come as soon as October 2026 [6]. Wikipedia's Anthropic entry, citing Reuters and Axios, records a subsequent $65 billion round in May 2026 at a reported $965 billion valuation, though we could not verify that against the original wire copy directly [7].

The strangest detail is the funder. Google, whose Gemini is Claude's most direct model competitor, is simultaneously one of Anthropic's largest backers and infrastructure suppliers. Whichever assistant wins the workload, Alphabet gets paid.

What do people actually use Claude for?

Coding, overwhelmingly. The Anthropic Economic Index found computer and mathematical tasks made up 37.2% of Claude.ai conversations in early 2025, and coding remained the largest category at 36% in the September 2025 report.

The first Anthropic Economic Index report, built on around one million Claude.ai conversations, found 37.2% of queries fell into the computer and mathematical category. Roughly 36% of jobs showed AI use for at least a quarter of their tasks, but only about 4% used it for three quarters or more [3].

The September 2025 report showed the mix shifting. Coding held at 36% of Claude.ai usage whilst education tasks rose from 9.3% to 12.4% and science from 6.3% to 7.2%. More strikingly, "directive" conversations, where the user hands the task over entirely, jumped from 27% in late 2024 to 39% by August 2025, the first time automation-style usage overtook augmentation [4].

Geography skews rich and technical. Anthropic's per-capita AI Usage Index put Israel top at 7.0x expected usage, then Singapore at 4.6x and Australia at 4.1x, and found a 1% rise in GDP per capita associated with a 0.7% rise in Claude use per capita [4]. India is now Claude's second-largest market at 5.8% of global usage, which is why TechCrunch reported Anthropic cutting India-specific pricing in July 2026 [8].

The honest caveat, which Anthropic itself makes, is that this data describes Claude's users, not the world's workers. Claude is marketed as a coding model, so coding-heavy usage is partly selection effect.

Is Claude really the enterprise and coding leader?

On the best available third-party measure, yes. Menlo Ventures' mid-2025 survey put Anthropic at 32% of enterprise LLM API usage, ahead of OpenAI at 25%, and at 42% of the code generation market specifically.

Menlo Ventures surveyed over 150 technical leaders in mid-2025 and found Anthropic leading enterprise API usage on 32%, with OpenAI on 25% (down from about 50% at the end of 2023) and Google on 20%. In code generation, Claude took 42% against OpenAI's 21%, and total model API spend had surged to $8.4 billion in six months [9].

Anthropic's own enterprise numbers point the same way. Customers spending over $100,000 a year grew nearly 7x in the year to February 2026, accounts worth $1 million-plus a year went from 12 to more than 500 in two years, and eight of the Fortune 10 are customers [1]. Business customers overall passed 300,000 by September 2025 [2].

Claude Code is the sharpest expression of this. It went from full launch in May 2025 to $500 million in run rate by that August [2], then past $2.5 billion by February 2026, with weekly active users doubling since 1 January and Claude models authoring around 4% of all public GitHub commits worldwide [1].

That last number deserves a pause. One vendor's agent is writing a measurable slice of the world's public code. It is Anthropic's own claim, unverified externally, but even half of it would be remarkable.

How many people use Google Gemini in 2026?

Google reported the Gemini app tripling its users year on year, passing 950 million monthly users by mid-2026 having stood at 650 million when Gemini 3 launched in November 2025. On iOS alone it was downloaded more than 137 million times in twelve months.

TechCrunch reported the trajectory from Alphabet's earnings: 650 million monthly users in late 2025, 750 million by February 2026, and past 950 million by Q2 2026, with iOS downloads topping 137 million over the trailing year [10]. Sundar Pichai credited features like Daily Brief and the Gemini Spark agent for the acceleration [10].

Behind the app sits a much larger machine. Alphabet's Q3 2025 commentary disclosed over 1.3 quadrillion tokens processed monthly across its surfaces, more than 20x annual growth, with Gemini app queries tripling quarter on quarter and AI Mode reaching 75 million daily users [11].

Comparisons with ChatGPT need care. OpenAI counts weekly users, Google counts monthly, and a Gemini "user" may simply be an Android owner who tapped a built-in button once. Monthly figures always look kinder than weekly ones.

What did Gemini 3 actually change?

Gemini 3, launched on 18 November 2025, moved Google to the top of the main model leaderboards and materially increased usage. Alphabet said Gemini 3 Pro processed three times as many daily tokens on average as its predecessor.

At launch, Google reported Gemini 3 Pro scoring 1501 Elo on LMArena, 76.2% on SWE-bench Verified and 91.9% on GPQA Diamond, with 13 million developers having built with its generative models [12].

The usage response was immediate. On the Q4 2025 earnings call, Alphabet said Gemini 3 Pro was processing three times as many daily tokens on average as 2.5 Pro, and that first-party models were handling over 10 billion tokens per minute through direct API use [13].

Benchmarks are marketing as much as measurement, and every lab times its wins. But the token multiplier is a behavioural signal rather than a leaderboard one: people did more with the new model, immediately.

How far does Gemini reach beyond the chatbot app?

Further than any rival. By early 2026 Gemini was embedded in Search, Workspace, Android, TVs and cars, with over 120,000 enterprises using it and AI Overviews at 2 billion monthly users.

Alphabet's Q4 2025 disclosures give the scale: more than 120,000 enterprises using Gemini, Gemini Enterprise passing 8 million paid seats across more than 2,800 companies, 95% of the top 20 SaaS companies building with it, and Circle to Search live on over 580 million Android devices [13].

Cars are the newest surface. TechCrunch reported Gemini rolling out from April 2026 to roughly 4 million General Motors vehicles from model year 2022 onward, with Google signalling it would not stay a GM exclusive [14].

Image generation shows how single features move the numbers. Google's Nano Banana editing model drove more than 10 million new users to the Gemini app and over 200 million image edits within weeks of its August 2025 release, figures Google announced and 9to5Google reported, as logged on the model's Wikipedia entry [15]. The Nano Banana-powered generation later spread into Gemini's Personal Intelligence features in 2026 [16].

This is the structural difference between the two companies. Anthropic has to sell Claude one developer and one enterprise at a time. Google can flip a default switch and reach a hundred million people by Friday.

Are Claude and Gemini catching ChatGPT?

On different axes, yes. ChatGPT still leads consumer scale, but Gemini is closing the user gap fast and Claude already leads enterprise API usage. The race stopped being one-dimensional in 2025.

ChatGPT reached a reported 1 billion monthly active users in June 2026, and Sensor Tower's H1 2026 data put its share of AI assistant usage below 50% for the first time, with Gemini at 27.7% [10]. Gemini's 950 million monthly users put it within touching distance of the leader.

Claude will never win that chart, and does not appear to be trying. Its 32% enterprise API share [9], $14 billion run rate [1] and Claude Code's grip on developers describe a company monetising a smaller audience far more deeply. Ofcom's UK research and Similarweb's traffic data consistently show Claude a distant third or fourth by consumer visits, which makes the revenue numbers all the more striking.

The 2026 pattern, then: ChatGPT owns the household name, Gemini owns distribution, Claude owns the invoice.

How to read these numbers

Almost every figure here comes from an interested party. Anthropic's revenue, customer and GitHub-commit numbers are self-reported in funding announcements designed to attract investors, and none are audited public accounts.

Alphabet's user metrics are selective earnings-call disclosures. "Monthly users" of a bundled assistant is a generous unit, and Google does not say how many Gemini users arrived via a default setting rather than a choice. Token counts measure machine activity, not human value.

Run rates annualise a recent period and assume nothing slows. Projections, such as Anthropic's Q2 2026 revenue and profit expectation, are exactly that. Third-party measures disagree with each other too: TechCrunch reported Anthropic's February round at a $350 billion valuation whilst Anthropic itself said $380 billion post-money, a gap explained by pre- versus post-money framing [1][6].

Survey-based market shares, like Menlo's, rest on around 150 respondents skewed towards large US enterprises and funded startups. Treat every number here as directional, and the direction as more trustworthy than the decimal.

What this means going into 2027

Enterprise AI spend is consolidating around Claude, consumer reach around Gemini. Menlo's data and Anthropic's customer growth suggest procurement teams have largely picked their coding model, whilst Google converts Android's installed base. Betting content or product strategy on ChatGPT alone now misses most of the market's growth.

Agentic, automation-style usage is becoming the default. With 77% of Claude's business API traffic showing automation patterns and directive consumer use rising fast [4], 2027's AI traffic will increasingly be software acting for people, not people typing. Design for the agent as well as the human.

Gemini's surfaces will keep multiplying. Cars, TVs, Workspace and Search mean Gemini touchpoints will exist for people who never download an AI app. Visibility in Google's AI answers is becoming inseparable from visibility in Google full stop.

Watch the profitability line, not the valuation line. Anthropic's flagged first profitable quarter matters more than its $380 billion sticker, because it would show frontier AI can fund itself. If compute costs push it back into loss, expect pricing across the industry to rise.

An Anthropic IPO would reset the scoreboard. Reporting points to a possible listing as early as October 2026 [6]. Public filings would replace curated run rates with audited numbers, and every statistic in this article gets marked to market.

If you want to work out which of these assistants your customers actually use, and how to be visible inside them, you can book a call with me and we will map it against your own numbers.