I've been using Claude since its first public release in March 2023. Since then, I've watched Anthropic grow at a pace that's unmatched even in the AI industry. From $10 million in revenue in 2022 to roughly $47 billion in annualized revenue in May 2026.
In this article, I've compiled the latest numbers, data and facts about Claude and Anthropic. From funding rounds and valuation to benchmarks and model development to enterprise adoption and safety research.
- Anthropic is valued at $965 billion after its Series H round (May 2026, post-money), having raised $65 billion. Annualized revenue crossed approximately $47 billion in May 2026 according to Sacra
- Claude Opus 4.5 was the first model to break the 80% mark on SWE-bench Verified. In February 2026, Claude Opus 4.6 became the first model to simultaneously hold the #1 position on all three LMSYS leaderboards
- Anthropic employs roughly 5,000 people (May 2026) and is a pioneer in Constitutional AI and AI safety research
1. Company Overview and Milestones
Anthropic was founded in 2021 by Dario and Daniela Amodei along with five other former OpenAI researchers. The reason for the split was concerns about AI safety and alignment at OpenAI.
Here's a quick overview of the numbers that define Anthropic's position in the AI industry:
1.2. Milestones from 2021 to 2026
In less than five years, Anthropic has evolved from a small research lab to one of the most valuable startups in the world:
In less than three years, Anthropic has evolved from a small research lab to one of the most valuable startups in the world. Pretty impressive, considering the company was founded by just seven people.
2. Funding and Investors
Anthropic's funding rounds are among the largest in tech history. In total, the company has raised over $119 billion:
2.1. Funding Rounds
The Series H at $65 billion (May 2026) is the largest private funding round of all time, surpassing even OpenAI's $40 billion round. Let that sink in for a moment.
2.2. Key Investors
Amazon and Google are the two largest strategic investors by a wide margin. The following chart shows the key backers by invested amount:
Amazon is the largest single investor with a total of $8 billion. In return, Claude is available on AWS Bedrock and Anthropic uses Amazon's Trainium chips for training. Google has invested a total of $3 billion and offers Claude via Google Cloud Vertex AI.
3. Valuation and Revenue
Anthropic's valuation has increased more than fifteenfold in just over a year. The following chart shows the exponential rise in post-money valuation over the past three years:
3.1. Revenue Growth
Anthropic's revenue growth is unprecedented. Since December 2024 alone, annualized revenue has grown from $1 billion to approximately $47 billion in May 2026:
Annualized Revenue (Run Rate)
That means growth from $1 billion to $47 billion in just 17 months. No technology company in history has grown this fast at this scale.
And yes, that's not a typo.
3.2. Revenue Comparison with Competitors
Anthropic overtook OpenAI in annualized revenue in April 2026. At $47 billion run rate, it now leads the field of frontier AI labs:
Annualized Revenue Comparison (May 2026)
Anthropic overtook OpenAI in annualized revenue in April 2026, confirmed by multiple independent sources. At $47 billion run rate (May 2026), Anthropic now leads OpenAI's $25 billion (February 2026).
4. Website Traffic and Usage
Monthly visits to claude.ai have grown significantly over the past year. In June 2026, claude.ai recorded approximately 946.7 million monthly visits according to SimilarWeb, a slight 0.6% dip from May's 952.6 million after climbing 16% from April and more than 5x year-over-year:
Monthly Visits to claude.ai
4.2. Claude in Traffic Comparison
Despite rapid growth, the gap to ChatGPT remains significant. But Claude is growing faster than any other competitor. In May 2026, Claude holds an 8.9% share of total AI chatbot traffic:
The shift becomes even clearer when you compare market share directly from April to May 2026. While ChatGPT slips from 54.7% to 52.7%, Claude is the only one of the four to gain, rising from 8.2% to 8.9%:
Additional usage data:
- Approx. 946.7M monthly visits (June 2026, down slightly from a 952.6M May peak, per SimilarWeb)
- 7.4M monthly active app users (December 2025)
- Over 300,000 business customers
- Approx. 70% of Fortune 100 companies use Claude
- Average session duration: 6:02 minutes per visit (May 2026)
- Bounce rate: 26.42% (May 2026)
4.1. Geographic Distribution
Claude is used worldwide, with the USA as the largest single market. Notably, India and Japan both rank among the top five countries:
Top Countries by Traffic Share
The highest per-capita usage is found in Singapore (4.6x above average), Canada (2.9x), Israel and Australia. In the US, Washington D.C. (3.82x) and Utah (3.78x) lead in per-capita usage.
5. Claude Models Overview
Anthropic has released numerous models since the launch of Claude 1.0 in March 2023. The following chart shows the knowledge cutoff dates of the individual models:
The current top Claude models offer up to 1 million tokens of context without the former beta surcharge. Claude Haiku 4.5 stays at 200,000 tokens. For a detailed comparison of all models, check out our article Claude Models Compared.
5.1. Model Overview
6. Benchmark Results
The bottom line: Claude is among the top-performing models in many benchmarks. Particularly in coding and reasoning, Anthropic regularly sets new standards.
6.1. Model Comparison on Key Benchmarks
| Benchmark | Fable 5 | Opus 4.8 | Opus 4.7 | Opus 4.6 | Opus 4.5 | Sonnet 4.6 | GPT-5 |
|---|---|---|---|---|---|---|---|
| SWE-bench | 95.0% | 88.6% | 87.6% | 80.8% | 80.9% | 79.6% | 74.9% |
| GPQA Diamond | -- | 93.6% | 94.2% | 91.3% | 87.0% | 89.9% | 85.7% |
| HumanEval | -- | -- | -- | 93.2% | 92.0% | 92.8% | 91.5% |
| MMLU | -- | -- | 92.4% | 91.1% | 90.5% | 90.4% | 90.8% |
| MGSM | -- | -- | -- | 92.0% | 91.6% | 91.2% | 90.0% |
Key highlights:
- Claude Opus 4.8 (May 2026) posts 88.6% on SWE-bench Verified, the best Opus-series score, ahead of Opus 4.7 (87.6%) and Opus 4.6 (80.8%); Fable 5 leads overall at 95.0%
- Claude Opus 4.8 reaches 93.6% on GPQA Diamond and 69.2% on the harder SWE-bench Pro
- Claude Opus 4.5 was the first model to break the 80% mark on SWE-bench Verified (80.9%)
- Claude Opus 4.6 held the #1 position on the LMSYS High-Elo Bracket with 1,503 Elo points (February 2026)
- Claude Opus 4.5 (with Extended Thinking) led the LMSYS Coding Arena with 1,510 Elo points (February 2026)
7. API Pricing Comparison
You might be thinking: If Claude performs this well, it must be more expensive, right?
Quite the opposite. Claude and GPT-5.5 are close in price. GPT-5.5 is actually cheaper on input ($2.50 vs. $3 for Sonnet 4.6), while both are tied on output at $15. Prices are per 1 million tokens:
On output, the picture flips. Here Opus 4.8 is the most expensive at $25 per 1 million tokens, while Sonnet 4.6 and GPT-5.5 are level at $15:
Additional discounts include Prompt Caching (90% cheaper on input price) and the Batch API (50% discount on input and output).
7.1. Subscription Pricing
8. Context Window and Technical Limits
Claude has one of the largest standard context windows among frontier models:
Standard Context Window Comparison
Since March 13, 2026, the 1 million token context window is generally available for Claude Opus 4.6 and Sonnet 4.6, with no beta header or price premium required. Claude can output up to 128,000 tokens per response (even more with Extended Thinking).
9. Claude Code
Claude Code deserves its own section. I use it daily and it has fundamentally changed my coding workflow.
9.1. Growth
Claude Code launched in February 2025 and became the most-used AI coding tool within just 8 months, overtaking both GitHub Copilot and Cursor. The numbers:
Claude Code's run rate in May 2026 is above $2.5 billion and has more than doubled since the start of the year. Business subscriptions for Claude Code have quadrupled since early 2026. Enterprise accounts now make up more than half of Claude Code's revenue.
10. Enterprise Adoption and Partnerships
Anthropic has formed strategic partnerships with some of the largest technology companies in the world. Enterprise customers now drive the majority of revenue growth.
10.1. Strategic Partnerships
10.2. Enterprise Market Share
The enterprise penetration is impressive. Approximately 70% of Fortune 100 companies use Claude. According to Ramp Economics Lab, Anthropic's share of combined enterprise spending on OpenAI and Anthropic grew from roughly 10% at the start of 2025 to over 65% in February 2026:
11. Constitutional AI and Safety
This is where Anthropic differs most clearly from the competition. While OpenAI and Google primarily focus on speed and new features, AI safety sits at the very top of Anthropic's agenda.
Constitutional AI is Anthropic's core approach. Instead of relying solely on human feedback (RLHF), Claude is trained based on a set of ethical principles. The model learns to evaluate and improve its own responses based on this "constitution."
Anthropic's safety work has also influenced legislation: California's SB 53, New York's RAISE Act and the EU AI Act (Codes of Practice) now require frontier AI developers to publish risk frameworks.
12. Team and Leadership
Anthropic was founded by seven former OpenAI employees. The leadership team includes:
- Dario Amodei (CEO): Previously VP of Research at OpenAI
- Daniela Amodei (President): Previously VP of Safety & Policy at OpenAI
- Jared Kaplan: Head of Science
- Jack Clark: Co-Founder, Head of Policy
- Chris Olah: Co-Founder, Interpretability Researcher
- Tom Brown: Co-Founder (Lead Author of GPT-3)
- Mike Krieger: Head of Product (Co-Founder of Instagram)
The team has grown significantly:
Employee Count Comparison (May 2026)
13. Claude vs. ChatGPT Comparison
I get this question all the time: Is Claude better than ChatGPT? In short, it depends on the use case. Here are the key metrics at a glance:
For a detailed comparison of the two AI assistants, check out our article ChatGPT vs. Claude.
13.1. Valuation Head-to-Head
Anthropic has now surpassed OpenAI in absolute valuation. At $965 billion (May 2026), Anthropic leads OpenAI's $852 billion (March 2026). On the revenue multiple (valuation divided by ARR), Anthropic at roughly 20x is actually below OpenAI's roughly 34x, meaning Anthropic is priced more conservatively relative to revenue:
Together, both companies are worth over $1.8 trillion on the private market, exceeding the market cap of companies like SAP and Siemens.
14. Facts and Records
And because statistics articles are simply better with a few entertaining facts:
- Namesake: Claude is named after Claude Shannon, the "father of information theory." Shannon's work "A Mathematical Theory of Communication" (1948) has been cited over 160,000 times.
- First Non-OpenAI Model #1: In March 2024, Claude 3 Opus became the first model from another provider to overtake GPT-4 on the LMSYS Chatbot Arena Leaderboard.
- Sibling Founders: Anthropic was founded by siblings Dario and Daniela Amodei. Both previously held VP positions at OpenAI.
- Fastest Revenue Growth: From $1 billion to $47 billion ARR in just 17 months (Dec 2024 to May 2026).
- SWE-bench Record: Claude Opus 4.5 was the first AI model to break the 80% mark on SWE-bench Verified (80.9%) in November 2025.
- Largest VC Round: The Series H at $65 billion in May 2026 is the largest venture capital round of all time.
- Claude Code Overtaking: Claude Code launched in February 2025 and overtook GitHub Copilot and Cursor within 8 months as the most-used AI coding tool.
- Deloitte Record: The deployment of Claude for 470,000+ Deloitte employees is the largest enterprise AI deployment by a single AI provider.
- Multi-Cloud Strategy: Anthropic is the only major AI company simultaneously using Google TPUs, Amazon Trainium chips and NVIDIA GPUs.
- Shannon as Investor: By the way, Claude Shannon was also a successful investor. He achieved an average return of about 28% per year from 1958 to 1986 (Warren Buffett: 27% from 1965 to 1995).
