Key Takeaways
- Anthropic has overtaken OpenAI. $30 billion ARR versus OpenAI’s $25 billion — the first time a rival has led OpenAI in revenue since ChatGPT launched in November 2022.
- 30× growth in 15 months. January 2025: $1B ARR. April 2026: $30B ARR. This is not a gradual trend — it is a near-vertical line on any chart.
- Enterprise won the race. Anthropic’s 80% enterprise revenue mix vs OpenAI’s consumer-heavy base explains the gap. 1,000+ companies each spending over $1 million annually. Claude Code holds 54% of the AI coding tool market.
- IPO October 2026. Anthropic is targeting an October IPO raising $60B+ at a $380B valuation — and now arrives at that listing with a stronger revenue position than OpenAI.
The Number That Changes Everything
On April 7, 2026, Anthropic announced its annualised revenue run-rate has surpassed $30 billion.
OpenAI’s figure, as of end of February 2026, was $25 billion.
For the first time since ChatGPT made generative AI a household term in late 2022, a rival has overtaken OpenAI in revenue. The company that was considered to have an insurmountable brand advantage, an unassailable user base, and a multi-year head start is now — at least by this metric — in second place.
The speed of the reversal is what makes this extraordinary.
Direct Answer: Has Anthropic overtaken OpenAI in revenue? Yes. As of April 7, 2026, Anthropic’s annualised revenue run-rate has surpassed $30 billion, compared to OpenAI’s $25 billion (as of end of February 2026). This is the first time Anthropic has led OpenAI in revenue. Note: the two companies use different revenue accounting methodologies, so the comparison is not perfectly apples-to-apples. But even accounting for methodological differences, Anthropic’s growth trajectory has clearly closed and crossed OpenAI’s position in the enterprise AI market.
The Growth Trajectory: 30× in 15 Months
| Date | Anthropic ARR |
|---|---|
| January 2025 | $1 billion |
| May 2025 | $3 billion |
| August 2025 | $5 billion |
| December 2025 | $9 billion |
| February 2026 | $14 billion |
| April 2026 | $30 billion |
Each step in this table is larger than the one before it — not just in absolute dollars but in time compression. The jump from $14B to $30B happened in approximately six weeks.
Dario Amodei, Anthropic’s CEO, has publicly acknowledged that he consistently underestimates his own company’s growth. “I never assume anything, and actually I am always very conservative in saying I think it’s going to slow down on the business side,” he said last year. He has been wrong every time.
Epoch AI had predicted Anthropic would overtake OpenAI by mid-2026. It happened in April.
What Is Driving It: Claude Code and Enterprise Focus
The revenue story is inseparable from two things: Claude Code and Anthropic’s deliberate choice to prioritise enterprise over consumers.
Claude Code: $2.5 Billion From One Product
Claude Code — Anthropic’s agentic command-line coding tool — generates $2.5 billion in annualised revenue as of February 2026. That number is almost certainly higher now. Weekly active users doubled since January 1. Business subscriptions to Claude Code quadrupled in the first quarter of 2026.
Claude Code holds 54% of the AI coding tool market — ahead of GitHub Copilot and Cursor. The leaked source code we covered last week revealed why: a 512,000-line, production-grade system with sophisticated permission architecture, multi-agent orchestration, and a three-layer memory system that competitors had not matched.
For context: $2.5 billion ARR from a single developer tool, in its first year at scale, is faster growth than Salesforce, Workday, or ServiceNow achieved at comparable stages.
Enterprise Mix: The Structural Advantage
Anthropic generates approximately 80% of revenue from enterprise customers — companies paying for API access and Claude integrations. OpenAI’s mix is more consumer-heavy, driven by ChatGPT subscriptions.
The enterprise mix creates compounding advantages:
Higher retention. Enterprise contracts renew on multi-year cycles. Consumer subscriptions churn monthly. Anthropic’s customer relationships are structurally stickier.
Expansion economics. When an enterprise customer deepens Claude integration into their workflows, their spend grows without Anthropic needing to re-acquire them. The 1,000 customers each spending $1M+ are the base of a much larger expansion curve.
Enterprise API share. In the LLM API market specifically — where companies buy model access to build products — Anthropic holds 32% share versus OpenAI’s 25%. Seven out of ten new enterprise customers are choosing Anthropic.
Fortune 10 penetration. Eight of the ten largest US companies by revenue are Anthropic customers. This represents institutional validation that is exceptionally difficult for competitors to displace.
The Broadcom-Google TPU Deal: Securing the Infrastructure
Alongside the revenue announcement, Anthropic revealed a major infrastructure expansion — a multi-gigawatt TPU deal with Google and Broadcom.
The deal gives Anthropic access to approximately 3.5 gigawatts of TPU-based computing capacity beginning in 2027. This is structured as an extension of Anthropic’s November 2025 commitment to invest $50 billion in American AI infrastructure.
The significance: Anthropic is now securing compute capacity years in advance based on its revenue trajectory. This is the behaviour of a company that has high confidence in sustained demand growth — not a company hedging against uncertainty.
Broadcom’s role is notable. The custom chip designer joins Google in this arrangement, potentially supplying purpose-built AI accelerators specifically designed for Anthropic’s inference workloads. Custom silicon optimised for Claude’s architecture would represent a structural cost advantage over competitors running on commodity GPU infrastructure.
The Pentagon Clash: A Sovereignty Story
Buried in the revenue announcements is a detail that deserves direct attention for Vucense readers.
Anthropic recently refused a Department of Defense demand to make Claude available for “all lawful purposes” — a requirement that would have included autonomous weapons systems and mass surveillance applications. The Pentagon responded by labelling Anthropic a “supply chain risk.”
Anthropic refused anyway. And then posted $30 billion in ARR.
The sequence matters. Anthropic’s largest customers are not the US military — they are enterprises building AI-powered products. The enterprise customers who chose Anthropic specifically because of its safety focus and refusal to compromise on use-case restrictions did not leave when the Pentagon clash happened. Revenue accelerated.
This is the first significant data point showing that a principled refusal to serve certain government use cases does not cost an AI company its commercial position. It may, for Anthropic’s specific customer base, reinforce it.
The IPO Race: Anthropic vs OpenAI
Anthropic is targeting an October 2026 IPO raising $60 billion+ at a $380 billion valuation. OpenAI is targeting a similar Q4 2026 window.
The race now looks different than it did six weeks ago when we covered it.
At that point, OpenAI had the higher revenue and the stronger brand. Anthropic had better fundamentals (higher capital efficiency, stronger enterprise mix) but lower absolute revenue.
Today: Anthropic has the higher revenue. The fundamentals advantage it had before remains. The question for the IPO is whether the market assigns Anthropic a premium for higher revenue + better fundamentals, or whether OpenAI’s 900 million weekly ChatGPT users command a consumer-brand premium that overrides the revenue gap.
The IPO that lists first will likely absorb significant pent-up retail demand for AI equity. Both companies know this. The October window is now a genuine race.
The Vucense Sovereignty Angle
Three things stand out from this story for readers who care about AI sovereignty:
Enterprise won over consumer. The revenue model that just overtook OpenAI is built on selling API access to companies building AI-powered products — not on surveillance advertising or consumer data monetisation. The $30B ARR is earned by being useful to businesses, not by harvesting user attention.
The Pentagon refusal held. Anthropic was labelled a supply chain risk for refusing to arm autonomous weapons with Claude. Revenue accelerated anyway. This is evidence that the market can sustain principled limits on AI use cases — at least at the enterprise level.
Claude Code is the product story. $2.5B ARR from a developer tool that runs locally, maintains sophisticated permission boundaries, and was built on a genuine engineering culture (512,000 lines of it, now public). The product that is driving Anthropic’s dominance is not a mass-market consumer app. It is a serious professional tool that serious developers are paying serious money for.
FAQ
How does Anthropic calculate its $30B ARR? Annualised revenue run-rate projects current monthly revenue forward over 12 months. Anthropic and OpenAI use different methodologies — Anthropic reports gross revenue including cloud computing credit arrangements, while OpenAI uses a net treatment. This means the comparison requires some caution, though the direction of the reversal is not in dispute.
Is Anthropic profitable? No. Despite $30B ARR, Anthropic is not yet profitable. Compute costs are enormous. The company projects positive free cash flow by 2027 and profitability by 2028 — significantly faster than OpenAI’s 2030 target.
What is Claude Code’s market share? 54% of the AI coding tool market as of early 2026, generating $2.5B ARR. The next largest competitors are GitHub Copilot and Cursor.
When is Anthropic’s IPO? Anthropic is targeting October 2026, aiming to raise $60B+ at a $380B valuation. No S-1 has been filed. The timeline is subject to market conditions and SEC review of accounting methodology questions.
Did the Pentagon labelling Anthropic a “supply chain risk” hurt business? Based on the revenue data: no. Revenue accelerated from $19B to $30B ARR in the weeks after the Pentagon clash became public. The enterprise customer base that drives Anthropic’s revenue appears to have either ignored or positively responded to Anthropic’s refusal to compromise.
Related Articles
- OpenAI vs Anthropic: The $1 Trillion IPO Race
- Anthropic Accidentally Leaked Claude Code’s Entire Source Code
- Was the Claude Code Leak Genius Marketing or a Real Mistake?
- Anthropic’s Mythos Model Tier: What We Know About Claude Beyond Opus
- OpenAI’s Ad Pilot Makes Claude’s No-Ads Promise Look Prophetic
Sources & Further Reading
- MIT Technology Review — AI Section — In-depth coverage of AI research and industry trends
- arXiv AI Papers — Pre-print research papers on AI and machine learning
- EFF on AI — Civil liberties perspective on AI policy
