Anthropic’s revenue growth and the accounting gap explained

Anthropic’s annualized revenue reached $65 billion, driven largely by Claude Code and a massive surge in enterprise customers. The company’s lead over OpenAI is partially attributed to gross revenue accounting methods used for cloud provider partnerships.

Anthropic's revenue growth and the accounting gap explained

Anthropic’s annualized revenue exceeded $65 billion this year. This massive figure follows an intense period of expansion from $9 billion at the end of 2025. In the first part of 2026, the company reported that its annualized revenue passed $47 billion. These numbers show a significant increase in how much businesses spend on artificial intelligence. The company added tens of billions of dollars in annualized revenue while it was still multiplying from a large base. This growth surpassed all other competitors in the AI sector. Claude Code drove much of this expansion. It brought developers into the ecosystem and then facilitated much larger enterprise accounts. The customer base of seven-figure spenders doubled to more than 1,000 in less than two months.

The logic behind the revenue numbers

The difference between Anthropic’s $65 billion revenue and OpenAI’s $40 billion revenue is not entirely due to market share. A significant part of this gap comes from different accounting methods. Anthropic records the full value of Claude sales through its partners and then records the cloud providers’ cuts as expenses, which allows the company to report much higher top-line revenue than competitors who use a net revenue accounting model. This method allows Anthropic to report its gross revenue because it acts as the principal in each transaction. If a customer pays $100 for an AI service through a cloud provider, Anthropic records the full $100 as top-line revenue. The cloud provider’s cut is then listed as an expense. OpenAI follows a different approach and records only its share of certain sales through partners like Microsoft. OpenAI sees net revenue as more accurate. Accounting is an art rather than a science. This accounting difference alone accounts for a significant part of the $25 billion gap. If Anthropic changed its accounting to match OpenAI, its revenue would drop by about 6% to 10%. This adjustment would still leave Anthropic $19 billion to $21 billion ahead of OpenAI.

The software engine driving growth

Claude Code is a major driver of this revenue. It is a developer tool that integrates via command-line and editor plugins like VS Code and JetBrains. The tool reached $1 billion in run-rate revenue six months after its public launch in May 2025. By February 2026, the company said Claude Code run-rate revenue exceeded $2.5 billion. This product uses a lot of computing capacity because developers use it for tasks like debugging, testing, and refactoring. This intensive usage leads to higher revenue per user than a standard chatbot. Claude Code accounts for 4% of all public GitHub commits and has 29 million daily installs on VS Code. The company also acquired Bun, a high-performance JavaScript runtime, to expand these capabilities. Claude Code uses Claude Skills, which are lightweight, YAML-fronted Markdown patterns that Claude loads on demand. These skills allow for specialized tasks like document creation and Slack-optimized GIF generation. Large companies like Air India and Cognizant use these tools to speed up software creation and manage DevOps workflows.

The enterprise and vertical markets

Most of the revenue comes from enterprise customers. Sacra estimates that more than 300,000 business customers accounted for about 80% of revenue in October 2025. These customers use API calls and enterprise subscriptions. This usage-based model means revenue can grow even if the company does not add new customers. As companies process more tokens, they pay more. The number of business customers spending over $1 million annually increased from 500 in February to more than 1,000 in under two months.

Anthropic has built specialized tools for different industries. Claude for Healthcare is a HIPAA-ready offering that includes native integrations to the CMS Coverage Database, ICD-10 codes, and PubMed. The company also acquired Coefficient Bio to add biological modeling capabilities to its healthcare unit. Claude for Finance provides connectors to LSEG, Moody’s, Aiera/Third Bridge, Chronograph, Egnyte, and MT Newswires. It also includes a beta Claude for Excel add-in for Max, Enterprise, and Teams users. Claude Design enables the creation of prototypes and slides with exports to Canva, PDF, PPTX, and standalone HTML. It also has design-system imports so outputs conform to company standards. Claude Cowork is an agent feature for the macOS desktop app that reads and manipulates files in user-selected folders. It runs in an isolated virtual machine. For legal needs, the company provides a suite of over a dozen plug-ins that automate document search and contract review.

Model tiers and token pricing

Anthropic provides different models for different needs. The pricing depends on whether a user needs speed or reasoning.

Plan Claude Code Claude Science Claude in Chrome
Free No No No
Pro Yes Yes Yes
Team Yes Yes Yes
Enterprise Yes Yes Yes
Model Input (per 1M tokens) Output (per 1M tokens)
Claude Opus 4.8 $5 $25
Claude Sonnet 4.6 $3 $15
Claude Haiku 4.5 $1 $5

Developers can reduce costs through specific features. The Batch API processes requests asynchronously within a 24-hour window and provides a 50% discount on all token costs. Prompt caching stores previously processed portions of a prompt and reduces repeated input costs by up to 90%. For high-volume tasks, the combination of these two features can reduce effective API spend by up to 95%.

The margin and efficiency jump

The company’s revenue growth is decoupling from its compute costs. In 2024, the blended gross margin was negative 94%. Today, it is in the mid-60% range, and API-only margins exceed 80%. This improvement comes from inference-stack optimization. The revenue per megawatt went from $16M to $60M in nine months. Anthropic also has high revenue per employee. Its revenue per employee is $9,013,869. This is higher than Nvidia’s $5,141,381 and much higher than Apple’s $2,506,994.

The capital race and IPO history

Anthropic is preparing for a massive IPO. It is targeting a valuation of $2 trillion or more in October. This would be the largest IPO in history. In May 2026, the company raised $65 billion at a $965 billion valuation. This round included investors like Altimeter Capital, Dragoneer Investment Group, and Sequoia Capital. In February 2026, it raised $30 billion at a $380 billion valuation. In January 2026, it raised $10 billion. In September 2025, the company raised $13 billion at a $183 billion valuation. In March 2025, it raised $3.5 billion at a $61.5 billion valuation. In January 2025, it raised $1 billion at a $60 billion valuation. These funds help the company secure compute resources and manage the high costs of training frontier models.

The competitive landscape

OpenAI remains a primary competitor. OpenAI is on track to exceed $40 billion in annualized revenue. However, OpenAI has a massive base of 900 million free users. These users cost money in compute power but do not pay for subscriptions. Anthropic targets enterprise customers from the start. This strategy avoids the high cost of maintaining a massive free user base. Does the massive difference in user types mean that the revenue comparison will never be truly accurate?

airtrain.ai
airtrain.ai

The airtrain.ai newsroom covers AI research, models and the tools built on them.

More on this topic

Stay ahead of AI

Get the week's most important AI stories delivered to your inbox every Monday.

No spam. Unsubscribe anytime.

More Stories