Last Updated: November 13, 2026

Claude Code Statistics 2026: Revenue, Users and Market Share at a Glance
Claude Code passed $2.5 billion in run-rate revenue in February 2026, the last Claude Code figure Anthropic has disclosed, and 39% of professional developers used it at work by mid-2026. The revenue number comes from Anthropic and the usage figure from JetBrains survey data. In November 2026, Menlo Ventures estimated that Claude Code and OpenAI Codex together take $24.0 billion, or 70%, of the $34.0 billion enterprises spend on AI coding. Neither Anthropic's S-1 nor its November amendment breaks out Claude Code revenue.
This page collects the Claude Code numbers that have a public source behind them: company announcements, Anthropic's S-1 filing, earnings calls, large developer surveys and major news outlets. Where a figure is an estimate or comes from a single secondary source, we say so.
Key Claude Code numbers (November 2026):
$2.5 billion+ run-rate revenue as of February 12, 2026, more than double the level at the start of the year (Anthropic)
About 18% of Anthropic's $14 billion run rate at that time (our calculation from the same Anthropic announcement)
$24.0 billion: enterprise spend on Claude Code and OpenAI Codex combined, 70% of a $34.0 billion enterprise AI coding market, in Menlo Ventures' 2026 estimate from a survey of nearly 400 US enterprise decision-makers (Menlo Ventures, November 2026)
39% of 15,000+ professional developers use Claude Code at work, up from 18% in January 2026; 47% in the US (ADTmag, I Programmer, both reporting JetBrains data)
46% of 906 engineers named Claude Code their most loved AI tool, versus 19% for Cursor and 9% for GitHub Copilot (The Pragmatic Engineer)
4% of public GitHub commits were authored by Claude Code in early February 2026 (Anthropic, citing SemiAnalysis)
About 20x: what the top 5% of Claude Enterprise users who use Claude Code spent in Q3 2026, compared with the under-$75 monthly average for Enterprise users without it (The New Stack, Alex Kras, both quoting the S-1)
$20/month for Claude Pro ($17/month billed annually), $100 for Max and $200 for Ultra, all with Claude Code included (Claude pricing)
Everything about Claude
Free download: The AI Stats Sheet 2026. 36 key AI numbers for business on four printable pages, with the source under every figure. Join 2,000+ business leaders who read AI Business Weekly and get it in your inbox right away.
Table of Contents
What Is Claude Code?
Claude Code is Anthropic's AI coding agent. Developers describe what they want in plain language, and Claude Code reads the codebase, writes and edits files, runs tests and commands, and keeps iterating until the task is done. It works across a whole project rather than suggesting one line at a time.
Anthropic released Claude Code as a research preview in February 2025, per its S-1, and made it generally available in May 2025. Unlike most AI coding tools, which started as IDE extensions, it began as a command-line tool that runs in the terminal, so it can plug into any editor, script or CI pipeline.
Claude Code runs on Anthropic's current models. The lineup listed on Anthropic's API pricing page in November 2026 is Claude Fable 5.2, Opus 5.5, Sonnet 5.5 and Haiku 5.5. Fable 5.2 arrived on November 9, 2026, and fast mode for Fable 5.2 is available in Claude Code, per Anthropic. Claude Code has a public GitHub repository, but it is not open source: it is distributed under Anthropic's commercial terms, as TechCrunch and OpenHands have both noted.
For the full breakdown of what Claude Code can do in practice, our what is Claude Code guide covers capabilities and deployment in detail. For hands-on walkthroughs, see our guides to Claude Code Remote Control and Claude Code plugins. To get more out of it day to day, our Claude Code best practices guide sums up what Anthropic recommends.
How Much Revenue Does Claude Code Make?
Claude Code's last company-reported revenue figure is a run rate of more than $2.5 billion on February 12, 2026, according to Anthropic's Series G announcement. That was up from $1 billion in November 2025 and $500 million in September 2025. Anthropic has not disclosed a Claude Code revenue figure since then.
Claude Code run-rate revenue milestones (company-reported):
Date | Milestone | Source |
|---|---|---|
February 2025 | Research preview launch | |
September 2025 | Over $500 million run rate, usage up 10x in three months | |
November 2025 | $1 billion run rate, about six months after general availability | |
February 12, 2026 | Over $2.5 billion run rate, more than doubled since January 1 | |
October 26, 2026 | S-1 filed; no standalone Claude Code revenue line |
Run rate is an annualized snapshot of current revenue, not reported annual sales. You will see higher Claude Code figures for mid-2026 on other sites, such as $8 billion. We could not trace those to Anthropic or a major outlet, so they are not in this table.
Revenue mix: Enterprise use made up more than half of Claude Code revenue by February 2026, and business subscriptions to Claude Code had quadrupled since the start of the year, per Anthropic. On the same date Anthropic's total run rate was $14 billion, so Claude Code was about 18% of company revenue (our calculation).
Anthropic's growth since then: Anthropic's run rate crossed $47 billion in early May 2026 (Anthropic Series H) and reached about $81 billion by the end of October, up from roughly $71 billion in late September (Fortune). Anthropic's amended S-1, filed November 2, reports Q3 2026 revenue of $17.0 billion and $33.2 billion for the first nine months of 2026 (Amendment No. 1 to the S-1; figures as tallied by sec-api.io). How much of that comes from Claude Code is not disclosed.
What the S-1 does say about Claude Code spending: In Q3 2026, Claude Enterprise users who did not use Claude Code spent less than $75 a month on average, while the top 5% of Enterprise users who did use it spent roughly 20 times that, per The New Stack and Alex Kras. Heavy coding users are among Anthropic's most valuable customers.
No neutral source measures Claude Code's share of AI coding spend on its own. The closest figure is Menlo Ventures' November 2026 estimate that Claude Code and OpenAI Codex together hold 70% of a $34.0 billion enterprise coding market. By usage, JetBrains survey data shows 39% of developers using Claude Code at work in mid-2026, ahead of GitHub Copilot at 21%.
The 2026 Menlo Ventures update: Menlo's 2026 enterprise report, based on a Morning Consult survey of nearly 400 US enterprise decision-makers fielded September 8 to 26, 2026, puts combined Claude Code and Codex spend at $24.0 billion. Coding is the largest enterprise AI application category at $34.0 billion, up 7.7x, and Anthropic's share of enterprise LLM API spend rose from 40% to 55% while OpenAI's rose from 27% to 31% (Menlo Ventures). Menlo's estimate and Anthropic's own February figure leave no public split of that $24.0 billion between Claude Code and Codex, so any single-product number you see is a guess.
The 54% figure, explained: Many sites describe Claude Code as holding "54% of the AI coding market." The original number comes from Menlo Ventures' December 2025 enterprise report (495 US enterprise AI decision-makers). It estimates Anthropic's models held 54% of enterprise coding spend in 2025, against 21% for OpenAI, and credits Claude Code for much of the gain. It measures model spend, not Claude Code seats or users.
The same Menlo Ventures report put Anthropic's share of enterprise LLM API spend at 40% in 2025, up from 24% in 2024 and 12% in 2023.
Developer survey data:
The Pragmatic Engineer's 2026 AI tooling survey collected 906 responses between January 27 and February 17, 2026. It is a newsletter reader survey, so it leans toward engaged, senior engineers:
Claude Code: 46% "most loved", the top result
Cursor: 19% most loved
GitHub Copilot: 9% most loved
71% of respondents who regularly use AI agents named Claude Code as a tool they use, versus 46% for GitHub Copilot and 39% for Cursor (multiple answers allowed)
The larger sample comes from JetBrains. Its 2026 survey of more than 15,000 professional developers found 39% use Claude Code at work, compared with 21% for GitHub Copilot, 16% for OpenAI Codex and 12% for Cursor, as reported by ADTmag and I Programmer.
GitHub commit share:
In February 2026, Anthropic cited an analysis estimating that 4% of all public GitHub commits were authored by Claude Code, double the share a month earlier. That analysis came from SemiAnalysis, which projected Claude Code would reach 20% or more of daily commits by the end of 2026 (GIGAZINE). We have not found an updated public measurement, so treat the 20% figure as a projection, not a result.
For broader context on how Claude Code sits within the AI coding tools landscape, our AI coding tools statistics guide covers GitHub Copilot, Cursor, and all major competitors with current market share data.
💡 Finding this helpful? Get bite-sized AI news and practical business insights like this delivered free every morning at 7 AM EST.
How Many Developers Use Claude Code?
Claude Code became the most-used AI coding tool at work in 2026. JetBrains data shows 39% of more than 15,000 professional developers using it at work in mid-2026, up from 18% in January, with 47% in the US. Anthropic also said weekly active Claude Code users doubled between January 1 and February 12, 2026, without giving a total.
Share of professional developers using each tool at work (JetBrains):
Tool | January 2026 | Mid-2026 |
|---|---|---|
Claude Code | 18% | 39% |
GitHub Copilot | 29% | 21% |
OpenAI Codex | About 3% | 16% |
Cursor | 18% | 12% |
Sources: JetBrains Research, as reported by ADTmag and I Programmer; January figures from JetBrains' January 2026 AI Pulse survey. The same survey found 90% of developers use AI coding agents at work at least weekly and 68% use them daily.
Anthropic's own usage: Claude wrote more than 90% of contributions to Anthropic's codebase in September 2026, and the average Anthropic engineer produced more than 45 times as much code per day in Q3 2026 as in 2024, according to the S-1 as reported by The New Stack and Alex Kras. The filing does not define "contributions" or say how much of that code gets human review.
How Good Is Claude Code? Performance and Benchmarks
Benchmarks score the models behind Claude Code, not the product itself. The best-known public figure is Claude Opus 4.5's 80.9% on SWE-bench Verified at its November 2025 launch, reported by ITPro and Neowin. Claude Code now runs on newer models, including Opus 5.5 and Sonnet 5.5.
Why we removed the old benchmark table:
Earlier versions of this page listed a 58% "tool score" for Claude Code, a 95% first-try correctness rate and other leaderboard claims. We could not confirm any of those with Anthropic or an independent leaderboard, so they are gone. SWE-bench Verified tests whether a model can fix real GitHub issues from open-source projects. Scores change with each model release, and vendors run them under different setups.
A more useful signal than any single benchmark is what customers spend. Anthropic's S-1 reports net revenue retention of about 900% as of September 30, 2026, meaning existing customers were spending far more than a year earlier (Anthropic S-1).
Why the terminal design matters: Most AI coding tools began as autocomplete inside an editor. Claude Code started in the terminal, so teams can run it inside any editor, CI/CD pipeline or scripted workflow. For large companies with mixed development setups, that flexibility is a practical reason to adopt it.
For detailed benchmark comparisons including GitHub Copilot, Cursor, and Windsurf, our AI coding tools statistics guide covers the complete performance landscape.
Is Claude Code Open Source? GitHub and npm Statistics
Claude Code's GitHub repository had 154,017 stars on October 31, 2026, per Gradually AI, up from 149,665 on October 7 (OSS Insight tracks the same repository). The repository is public, but Claude Code is not open source. It ships under Anthropic's commercial terms, and Anthropic has sent takedown notices over attempts to reverse-engineer it (TechCrunch).
Claude Code GitHub and npm metrics (October 2026):
Metric | Figure | Source |
|---|---|---|
GitHub stars | 154,017 (October 31, 2026); 149,665 on October 7 | |
GitHub forks | 26,643 | OSS Insight (single source) |
npm downloads, 30 days | 52.9 million (Sept 7 to Oct 6, 2026) | Gradually AI, from npm's download API (single source) |
Share of public GitHub commits | About 4% (February 2026) | Anthropic, citing SemiAnalysis |
Projected commit share, end of 2026 | 20%+ of daily commits (projection, not a measurement) |
Reading the npm number: Download counts include build servers, mirrors and automated installs, so 52.9 million downloads in a month is not 52.9 million people. It is still a useful sign of how often Claude Code gets installed into real development environments.
Which Companies Use Claude Code?
Enterprise use made up more than half of Claude Code revenue by February 2026, and business subscriptions had quadrupled since January 1, per Anthropic. Eight of the Fortune 10 are Claude customers, and the S-1 shows Claude Code users are among the biggest spenders on Claude Enterprise.
Enterprise customer metrics:
About 20x spend: the top 5% of Claude Enterprise users who use Claude Code spent roughly 20 times the under-$75 monthly average of Enterprise users who do not, in Q3 2026 (The New Stack, Alex Kras)
11,900+ customers paid Anthropic $100,000 or more over the 12 months to September 30, 2026, up from more than 950 a year earlier (Anthropic S-1, The New Stack). This covers all Claude products, not only Claude Code.
Business subscriptions to Claude Code quadrupled between January 1 and February 12, 2026 (Anthropic)
500+ customers spending over $1 million a year on Claude in February 2026 (Anthropic), which the company later said had doubled to more than 1,000 (VentureBeat). Again, this is all Claude spending.
8 of the Fortune 10 are Claude customers (Anthropic)
Landmark enterprise deployments:
Anthropic has named Netflix, Spotify, KPMG, L'Oréal and Salesforce as Claude Code users (Anthropic). Its S-1 adds that NASA engineers used Claude Code to plan and write navigation commands for the Mars Perseverance rover, and that non-technical teams are increasingly adopting it (Anthropic S-1).
The compute signal: In May 2026, Anthropic agreed to use all of Colossus 1, the Memphis data center with more than 220,000 Nvidia GPUs and over 300 megawatts of capacity, as reported by VentureBeat and Forbes. Capacity of that size only makes sense if usage, much of it agentic coding, keeps climbing.
Beyond developers: Claude Cowork applies the same agent design as Claude Code to knowledge work in general. Anthropic's S-1 says Cowork's adoption has outpaced what Claude Code achieved at the same point after its February 2025 research preview (Anthropic S-1). For wider context, see our AI adoption statistics guide.
How Does Claude Code Compare With Copilot, Codex and Cursor?
Claude Code leads workplace usage in 2026, but the field is crowded. JetBrains data puts Claude Code at 39% of developers at work, GitHub Copilot at 21%, OpenAI Codex at 16% and Cursor at 12%. Codex is the fastest-rising rival, Copilot is pivoting to multi-agent orchestration, and Cursor now belongs to SpaceX.
Tool | Used at work (JetBrains, mid-2026) | Other latest data point |
|---|---|---|
Claude Code | 39% (up from 18% in January) | $2.5B+ run rate (Feb 2026); 46% "most loved" in Pragmatic Engineer survey |
GitHub Copilot | 21% (down from 29%) | 4.7M+ paid Copilot subscribers, up 75% year over year (Jan 2026) |
OpenAI Codex | 16% (up from about 3%) | 5M+ weekly active users (June 2026, per OpenAI) |
Cursor | 12% (down from 18%) | Acquired by SpaceX in an all-stock deal valued at about $60B (closed August 2026) |
Sources: ADTmag (JetBrains data), The Pragmatic Engineer, Microsoft FY26 Q2 earnings call, Constellation Research (OpenAI figure), GIGAZINE and Techzine (Cursor deal)
GitHub Copilot's response: On Microsoft's October 28, 2026 earnings call, Satya Nadella said agent-generated pull requests on GitHub rose 6x in six months and that agents now open more than a third of all PRs on the platform. He named Claude Code and OpenAI Codex among the agents working on GitHub, and said that "if you're a coding agent user, you want, in fact, Anthropic and OpenAI" (call transcript, Microsoft Investor Relations). Microsoft is positioning Copilot as the layer that orchestrates many agents, including Anthropic's.
Cursor's position: SpaceX agreed to buy Cursor maker Anysphere on June 16, 2026, and Cursor announced the deal was complete on August 14, with an all-stock value of about $60 billion (GIGAZINE, Techzine). Our news story on the deal has the details. Cursor's workplace usage in the JetBrains data fell from 18% to 12% over roughly the same period.
OpenAI Codex: Codex is the rival gaining ground fastest. OpenAI said Codex had more than 5 million weekly active users in June 2026 (Constellation Research), and JetBrains data shows its workplace usage rising from about 3% to 16%, now ahead of Cursor (I Programmer). Anthropic does not publish a comparable Claude Code user count, so a direct head-to-head on users is not possible.
For head-to-head comparisons, see our Claude Code vs Cursor guide, our GitHub Copilot statistics and our Cursor AI statistics. For the wider Anthropic picture, our Claude AI statistics guide covers the full platform.
What Changed for Claude Code in October and November 2026?
October and November 2026 brought the most detailed look yet at Anthropic's business. Anthropic filed its public S-1 on October 26 and an amendment with Q3 results on November 2, Menlo Ventures put Claude Code and Codex at 70% of enterprise coding spend, and Anthropic set its IPO price range on November 10. What did not arrive was a fresh Claude Code revenue number.
October 26: Anthropic filed its public S-1, reporting $16.3 billion in first-half revenue, more than 11,900 customers paying $100,000 or more and net revenue retention of about 900%.
Claude Code in the S-1: The filing calls Claude Code the product that has "most concretely defined what autonomous AI work looks like," and gives the 20x Enterprise spend gap and the 90% codebase contribution figure covered above (The New Stack).
October 28: Microsoft said agent-generated pull requests on GitHub rose 6x in six months, naming Claude Code among the agents (transcript).
November 2: Anthropic's amended S-1 reported Q3 2026 revenue of $17.0 billion, up 46.7% from Q2, and more than 1,900 customers above $1 million (S-1/A, sec-api.io). It still gives no Claude Code revenue line.
November 3: Menlo Ventures estimated Claude Code and Codex at $24.0 billion, 70% of the $34.0 billion enterprise coding market (Menlo Ventures).
November 9: Anthropic released Claude Fable 5.2 at $8/$40 per million tokens, 20% below Fable 5.1, with fast mode available in Claude Code (Anthropic).
November 10: Anthropic set an IPO price range of $70 to $78 a share for 825 million shares on the Nasdaq under the ticker ANT, with trading possible as early as the following week (Fortune).
Neither the S-1 nor its November amendment reports a standalone revenue figure for Claude Code, so any 2026 Claude Code revenue number after February is an outside estimate, not a company disclosure.
How Much Does Claude Code Cost in 2026?
Claude Code is included in every paid Claude plan. Claude Pro costs $20 a month, or $17 a month billed annually. Max costs $100 a month and Ultra $200 a month, per Claude's pricing page. Teams pay $25 per standard seat or $125 per premium seat each month, and API use is billed per token.
Plan | Monthly price | Billed annually |
|---|---|---|
Pro | $20 | $17/month ($200 up front) |
Max | $100 | Monthly only |
Ultra | $200 | Monthly only |
Team, standard seat | $25 per seat | $20 per seat |
Team, premium seat | $125 per seat | $100 per seat |
Source: claude.com/pricing, checked November 12, 2026. Max, Ultra and premium Team seats can use Fable for up to 50% of weekly limits, while Pro reaches Fable through usage credits. Ultra sits at the same $200 price point as the top Max tier Anthropic launched in 2025 (Engadget).
API prices for the models behind Claude Code (per million tokens):
Model | Input | Output |
|---|---|---|
Claude Fable 5.2 | $8 | $40 |
Claude Opus 5.5 | $4 | $20 |
Claude Sonnet 5.5 | $2 | $10 |
Claude Haiku 5.5 | $0.10 (prompts up to 100,000 tokens; $0.50 above) | $0.50 |
Source: Claude API pricing. Our Claude API pricing guide covers caching and batch discounts.
What Do These Numbers Mean for Engineering Teams?
Three separate signals point the same way. Claude Code's run rate went from $500 million to more than $2.5 billion in about five months, per Anthropic. Surveys from JetBrains and The Pragmatic Engineer both put it first. And the S-1 shows its heaviest Enterprise users spend about 20 times more than other Enterprise users.
Each signal has limits. Survey samples lean toward engaged developers, run rates are snapshots, and Anthropic has not published Claude Code revenue since February. Read any single number on this page as a direction, not a precise measurement.
For CTOs and engineering leaders evaluating Claude Code:
Budget for usage, not just seats. The S-1 spend data shows heavy Claude Code users cost far more than a typical Enterprise seat, so per-seat pricing alone will understate the bill for agent-heavy teams. Run a pilot with usage reporting turned on before a wide rollout.
The task mix matters. Coding agents do best on well-defined work: bug fixes, tests, migrations and features in familiar code. Novel architecture and vague requirements still need closer review from senior engineers.
Measure your own results. In a 2025 randomized study, METR found experienced open-source developers took 19% longer on tasks when allowed to use early-2025 AI tools, even though they expected to be faster (ITPro, Reuters via Yahoo). The tools have improved a lot since then, but the lesson holds: track cycle time and defect rates for your team instead of trusting vendor claims.
For practical guidance on deploying AI coding tools across an engineering organization, our AI for business guide covers implementation frameworks including change management and measurement.
What is Claude Code? Complete Guide 2026
How Claude Code works, what you can build with it, and how to get started.
AI Coding Tools Statistics 2026
Market data on GitHub Copilot, Cursor, Claude Code, Codex and Windsurf, compared.
Claude AI Statistics 2026
Anthropic and Claude platform data: revenue, users, valuation and enterprise adoption.
Anthropic Statistics 2026
Full Anthropic company data, including Q3 revenue, the IPO price range and S-1 figures.
GitHub Copilot vs Codeium vs Tabnine
Head-to-head comparison of AI coding tools for teams making deployment decisions.
AI Productivity Statistics 2026
How AI coding tools fit into the broader productivity data landscape including ROI benchmarks.
ChatGPT vs Claude: Which AI Is Better for Business?
How Claude's enterprise dominance and coding leadership compares to ChatGPT's consumer scale.
Get the next update first. We update these Claude Code numbers as new data comes out. AI Business Weekly sends the AI news and numbers that matter for business every weekday, free, and new subscribers can grab the free AI Stats Sheet 2026.
Frequently Asked Questions
What is Claude Code's revenue in 2026?
Anthropic's latest disclosed figure is a Claude Code run rate above $2.5 billion on February 12, 2026, up from $1 billion in November 2025, and enterprise use made up more than half of it (Anthropic). Anthropic's October 2026 S-1 does not break out Claude Code revenue, so later figures such as $8 billion are outside estimates, not company numbers.
What is Claude Code's market share?
No neutral source measures it on its own. Menlo Ventures estimated in November 2026 that Claude Code and OpenAI Codex together hold 70% of a $34.0 billion enterprise AI coding market (Menlo Ventures). JetBrains data shows 39% of developers used Claude Code at work in mid-2026, ahead of GitHub Copilot at 21%, OpenAI Codex at 16% and Cursor at 12% (ADTmag).
How many people use Claude Code?
Anthropic has not published a total user count. It said weekly active users doubled between January 1 and February 12, 2026 (Anthropic). JetBrains data shows 39% of more than 15,000 professional developers use it at work (I Programmer), and its npm package was downloaded 52.9 million times in the 30 days to October 6, 2026, a count that includes automated installs (Gradually AI).
How much does Claude Code cost?
Claude Code comes with every paid Claude plan: Pro at $20 a month ($17 billed annually), Max at $100 and Ultra at $200. Team seats cost $25 (standard) or $125 (premium) per month (Claude pricing). On the API, Sonnet 5.5 costs $2 per million input tokens and $10 per million output tokens, and Opus 5.5 costs $4 and $20 (Claude API pricing).
Is Claude Code better than GitHub Copilot?
It depends on what you measure. In The Pragmatic Engineer's 2026 survey, 46% of engineers named Claude Code their most loved tool versus 9% for Copilot (The Pragmatic Engineer), and JetBrains data shows more developers using Claude Code at work, 39% versus 21%. Copilot still had more than 4.7 million paid subscribers in January 2026 (Microsoft) and Microsoft's distribution behind it.
Is Claude Code better than Cursor?
Claude Code leads on survey preference (46% versus 19% most loved in The Pragmatic Engineer's survey) and on workplace usage (39% versus 12% in JetBrains data, per ADTmag). Cursor is an AI-native code editor rather than a terminal agent, and SpaceX has owned it since August 2026. Many teams use both. Our Claude Code vs Cursor guide compares them in detail.
Does Anthropic's S-1 include Claude Code numbers?
Some. It reports that the top 5% of Claude Enterprise users who use Claude Code spent about 20 times the under-$75 monthly average of Enterprise users who do not, and that Claude wrote over 90% of contributions to Anthropic's own codebase in September 2026 (The New Stack). Neither the S-1 nor the November 2 amendment gives Claude Code revenue, users or share of company revenue.
Is Claude Code open source?
No. Claude Code has a public GitHub repository with roughly 150,000 stars (OSS Insight), but the software is distributed under Anthropic's commercial terms, not an open-source license (TechCrunch).
Quick Answers
Who leads AI coding tools in 2026?
Claude Code, by workplace usage: 39% of developers in JetBrains data, ahead of GitHub Copilot at 21% (ADTmag).
How much revenue does Claude Code generate?
More than $2.5 billion in run-rate revenue as of February 12, 2026, the last figure Anthropic disclosed (Anthropic).
What share of Anthropic's revenue comes from Claude Code?
About 18% in February 2026, based on Anthropic's own $2.5 billion and $14 billion run-rate figures (Anthropic; our calculation). No newer split has been published.
When did Claude Code launch?
As a research preview in February 2025, with general availability in May 2025. It reached a $1 billion run rate in November 2025 (Anthropic).
What share of GitHub commits does Claude Code write?
About 4% of public commits in February 2026, with SemiAnalysis projecting 20% or more of daily commits by the end of 2026 (Anthropic, GIGAZINE).
Conclusion
Claude Code went from research preview to a $2.5 billion run rate in about a year, and by mid-2026 it was the AI coding tool developers used most at work, according to JetBrains data. Few developer tools have grown that quickly.
The gap in the data is revenue. Anthropic stopped giving Claude Code figures after February, and the S-1 folds Claude Code into the wider business. What the filing does show is that heavy Claude Code users are among Anthropic's highest-spending customers.
Competition is getting tougher. OpenAI Codex is climbing fast, GitHub is turning Copilot into a hub for many agents, and Cursor now sits inside SpaceX. For engineering leaders, the practical move is to pilot with real usage tracking and judge tools on your own codebase.
We will update this page after Anthropic's IPO, which could start trading as early as the week of November 16, 2026, or when the company publishes new Claude Code figures.
How we source this page: every figure links to a company announcement, filing, earnings call, survey publisher or news outlet. Where only one secondary source exists, we say so. In this update we removed figures we could not trace, including a mid-2026 Claude Code run rate of $8 billion, a 15,000-person sample wrongly attributed to The Pragmatic Engineer's survey, and several benchmark claims.
⭐ Get more AI Business Weekly in Google: add us as a preferred source on Google.
Written by Sameer Khan with AI assistance, fact-checked against the sources linked above.
