Anthropic Hit $19B in Revenue. Claude Code Alone Makes $2.5B. OpenAI Should Be Terrified.

Anthropic went from $9 billion in annualized revenue at the end of 2025 to $19 billion by March 2026. That's $10 billion added in roughly three months. CEO Dario Amodei confirmed the figure at a Morgan Stanley TMT conference, and the number is still climbing — Bloomberg reported the company was approaching $20 billion in run rate as of the first week of March.

To put this in context: OpenAI sits at approximately $25 billion in annualized revenue. The gap between the two companies has shrunk from a 10x advantage to less than 1.4x. Epoch AI's analysis projects Anthropic could overtake OpenAI by August 2026, with a 90% confidence interval spanning February 2026 to April 2027. Their internal projections are more conservative — Anthropic targeting 4x growth and OpenAI expecting 2.2x — but even those numbers suggest a crossover sometime in 2026 or 2027.

Claude Code Is the Growth Engine Nobody Predicted

The most striking data point buried in the revenue surge: Claude Code, Anthropic's terminal-based coding agent launched in mid-2025, now generates more than $2.5 billion in annualized billings. That's up from zero nine months ago, and the figure has doubled since January 1, 2026. Business subscriptions quadrupled in the six weeks after the new year.

Claude Code isn't a chatbot with a code editor bolted on. It runs inside your terminal, understands your full codebase through file system access, executes shell commands, manages git workflows, and can refactor entire modules autonomously. Developers using it report spending 60-80% less time on routine coding tasks — from writing tests to debugging production issues to handling pull request reviews.

The product has created a flywheel effect: developers who try Claude Code for a week rarely go back. The retention numbers, while not publicly disclosed, are reflected in that revenue doubling — you don't get that trajectory from free trials and churn.

Enterprise Dominance: 80% of Revenue, 500+ Million-Dollar Customers

Anthropic's revenue composition looks fundamentally different from OpenAI's. Over 80% of Anthropic's revenue comes from enterprise customers and API usage, compared to OpenAI's heavier reliance on consumer subscriptions through ChatGPT's 920 million users.

The enterprise numbers are staggering:

• 500+ customers now spend over $1 million annually — up from roughly a dozen two years ago
• 7x growth in customers spending $100,000+ per year
• 8 of the Fortune 10 companies are Claude customers
• 79% of OpenAI's paying customers also pay for Anthropic — enterprises are hedging their bets

That last statistic is perhaps the most revealing. Enterprises aren't choosing between Claude and ChatGPT — they're buying both. But when usage patterns shift and one model consistently outperforms on coding, analysis, and complex reasoning tasks, the budget allocation follows.

The $380 Billion Valuation and $30 Billion War Chest

In February 2026, Anthropic closed a $30 billion Series G round at a $380 billion post-money valuation. This makes it one of the most valuable private companies in history. The funding arrived at a critical moment — not because Anthropic needs cash to survive, but because the AI infrastructure buildout demands enormous capital.

Microsoft alone spends nearly $500 million annually on Anthropic services. Across the industry, hyperscalers — Amazon, Microsoft, Alphabet, and Meta — plan to spend approximately $650 billion on AI capital expenditure in 2026, a 71% increase from the $381 billion spent in 2025. Roughly 75% of that spending, or $450 billion, targets AI-specific infrastructure.

These companies are spending 90% of their operating cash flow on capex, up from 65% in 2025. Morgan Stanley expects hyperscaler borrowing to exceed $400 billion this year, more than double 2025's $165 billion. The AI infrastructure boom is real, and Anthropic is one of its primary beneficiaries.

The Pentagon Problem: Supply Chain Risk Designation

Anthropic's growth story has a significant complication. In March 2026, Defense Secretary Pete Hegseth formally designated Anthropic as a "supply chain risk" to national security, barring military contractors from conducting commercial activity with the company.

The conflict stems from Anthropic's two non-negotiable positions: opposition to AI being used in autonomous weapons systems, and opposition to AI-powered mass surveillance of U.S. citizens. The Pentagon said it could not accept these redlines. President Trump subsequently directed all federal agencies to cease use of Anthropic products within six months.

Anthropic CEO Dario Amodei responded that the company "does not believe this action is legally sound" and would challenge it in court. The company also argued the designation is narrower than Hegseth implied — it doesn't restrict uses of Claude or business relationships unrelated to Department of Defense contracts.

Despite the government crackdown, Anthropic's revenue trajectory hasn't flinched. The enterprise market clearly doesn't share the Pentagon's concerns about a company that refuses to build autonomous weapons.

The Growth Math: 10x vs 3.4x Per Year

Here's where the numbers get uncomfortable for OpenAI. Since reaching $1 billion in annualized revenue, Anthropic has grown at approximately 10x per year. OpenAI's growth rate over the same period has been 3.4x. Even accounting for a slowdown Anthropic experienced after July 2025 — where growth decelerated to 7x per year — the trajectory gap remains massive.

Epoch AI's piecewise exponential model, which accounts for this slowdown, still projects a crossover by December 2026 in the conservative scenario. The model's R-squared values are 0.99 for OpenAI and 0.98 for Anthropic, indicating high confidence in the trend lines.

The revenue per employee metric tells an equally dramatic story. Anthropic generates significantly more revenue per head than OpenAI, reflecting a leaner organization with higher leverage on its model capabilities.

What's Actually Driving the Divergence

Three factors explain why Anthropic is closing the gap:

1. Claude Code created a new market category. Before Claude Code, AI coding assistance meant autocomplete suggestions in your IDE. Claude Code turned the terminal into an autonomous programming partner. The $2.5 billion revenue line proves enterprise developers will pay premium prices for tools that genuinely reduce engineering time.

2. Enterprise-first strategy pays higher margins. API and enterprise contracts generate higher revenue per customer and are stickier than consumer subscriptions. When 500+ companies spend $1 million+ annually, the revenue base becomes highly predictable.

3. Model quality on technical tasks. Claude consistently ranks at or near the top on coding benchmarks like SWE-Bench, complex reasoning evaluations, and long-context tasks. For enterprises building AI-powered products, model quality on these specific capabilities drives purchasing decisions more than general chatbot fluency.

The Path to $43 Billion

Epoch AI's central projection shows Anthropic and OpenAI converging at approximately $43 billion in annualized revenue around August 2026. Whether that crossover actually happens depends on several variables: Claude Code retention rates, enterprise contract renewals, the Pentagon designation's legal outcome, and whether OpenAI's consumer business can sustain its growth rate.

But the trajectory is clear. In 18 months, Anthropic went from a distant second to a genuine threat to overtake the company that started the generative AI revolution. The $19 billion figure isn't a destination — it's a waypoint on a curve that shows no sign of flattening.