Anthropic Passes OpenAI in Revenue: $30B ARR vs $25B — 30x Growth in 15 Months and October IPO
In April 2026, Anthropic crosses $30B in annualized revenue, overtaking OpenAI for the first time since ChatGPT launched. From $9B to $30B in four months. IPO targeted for October at a $380B valuation.
On April 7, 2026, Anthropic announced that its ARR (Annualized Revenue Run-rate) has crossed $30 billion. OpenAI stands at $25 billion. For the first time since ChatGPT launched in late 2022, a direct competitor has pulled ahead. And the trajectory is nearly vertical: $1 billion in January 2025, $9 billion in December 2025, $14 billion in February 2026, $30 billion in April. That's 30x in 15 months. The most remarkable part: Anthropic is achieving this while spending four times less than OpenAI on model training. And the IPO is targeted for October 2026 at a $380 billion valuation.
The Steepest Curve in SaaS History
Anthropic's progression has no precedent in B2B software. Even OpenAI, which held the SaaS growth record, went from $1 billion to $10 billion in 18 months. Anthropic did it in 11. Then tripled again in 4 months.
| Date | Anthropic ARR | Multiple vs Early 2025 |
|---|---|---|
| January 2025 | $1B | 1x |
| June 2025 | $4B | 4x |
| December 2025 | $9B | 9x |
| February 2026 | $14B | 14x |
| April 2026 | $30B | 30x |
The key signal isn't the overall number. It's enterprise concentration. The number of Anthropic customers spending over $1 million annually grew from 500 to over 1,000 in less than two months, right after the Series G. Enterprises that prototyped Claude in 2025 moved to production in 2026. And when a Fortune 500 goes to production, the invoice stops being in tens of thousands — it becomes millions.
Why Anthropic Is Passing OpenAI Now
Three structural reasons explain the overtake.
Claude has become the enterprise coding standard. Claude Code, Cursor, and development agents run massively on Claude Sonnet 4.6 and Opus 4.6. Anthropic captures 42% of the coding assistant API market per internal data. OpenAI sits at 31%. That shift happened in less than 12 months.
Anthropic's enterprise pricing is more aggressive. Claude Sonnet 4.6 at $3/1M input tokens and $15/1M output beats GPT-4.5 at $5/$15 on mixed workloads. For an ISV doing 10 billion tokens per month, the difference is measured in millions.
Enterprise trust. Since launching Project Glasswing and Claude Mythos, Anthropic has positioned safety as a differentiator. Banks, insurers, and law firms prefer a model whose safety policy is public over a model whose product strategy shifts every quarter.
4x Less Compute for Higher ARR
The financial angle is even more striking. SaaStr calculated that Anthropic's 2025 training budget was around $5 billion, versus approximately $20 billion for OpenAI. Four times less. And Anthropic now posts $30 billion ARR versus $25 billion.
The CapEx-to-ARR ratio is emerging as the new benchmark of capital efficiency for AI labs. Anthropic hits 6:1. OpenAI sits at 1.25:1 — it burns almost as much as it earns. Investors haven't missed the signal.
| Metric | Anthropic | OpenAI |
|---|---|---|
| April 2026 ARR | $30B | $25B |
| 2025 Training Budget | ~$5B | ~$20B |
| ARR / Training Ratio | 6:1 | 1.25:1 |
| Customers > $1M/year | 1,000+ | Not disclosed |
| Recent Valuation | $380B (IPO target) | $852B (March post-money) |
OpenAI remains twice as valuable. But the ratio logic flips. At $852B on $25B ARR, OpenAI trades at 34x. Anthropic at $380B on $30B would trade at 12.6x. The multiple gap is starting to be justified by operational health.
October 2026 IPO: First Major Pure-Play AI Public Offering
Anthropic is targeting a Q4 2026 public offering. Target: raise $60 billion at a $380 billion valuation. It would be the largest tech IPO since Saudi Aramco in 2019.
Several signals converge. Reported bankers are Morgan Stanley and Goldman Sachs. Dario Amodei reportedly met with the SEC in March. And Anthropic recruited an experienced CFO out of Oracle in February.
The timing is strategic. Anthropic wants to be the first major pure-play AI public company, before OpenAI finishes its restructuring into a standard for-profit entity. Comparable public equities don't exist: Microsoft via OpenAI, Google via DeepMind, xAI not listed. Anthropic would be the only pure public exposure to the frontier layer.
OpenAI Isn't Dead — It Just Won't Dominate Alone
Some nuance is required. OpenAI remains bigger in consumer (ChatGPT at 700 million weekly users) and bigger in brand awareness (ChatGPT has become a common noun). Spud / GPT-6 ships between April 14 and May 5, 2026 and could flip the ARR trajectory.
But the enterprise market has become a duopoly — and Anthropic is taking the lead. Customers who bought OpenAI by default in 2024 tested Claude in 2025 and switched in 2026. Microsoft itself is multi-homing models on Azure — Claude Sonnet 4.6 has been available there since February. The era when OpenAI was synonymous with AI is over.
| Segment | Leader 2024 | Leader 2026 |
|---|---|---|
| Consumer (ChatGPT) | OpenAI | OpenAI |
| Enterprise API | OpenAI | Anthropic |
| Coding assistants | OpenAI | Anthropic |
| Cybersecurity | - | Anthropic (Mythos) |
| Brand awareness | OpenAI | OpenAI |
What It Changes for the Ecosystem
For startups building on the API. Pricing will stabilize: two credible players impose a discipline that a monopoly does not. Margins for ISVs selling AI agents should improve.
For enterprises making infrastructure choices. The "nobody gets fired for picking the leader" logic is cracking. CTOs can now justify an Anthropic choice by enterprise track record.
For investors. The Anthropic case proves that a disciplined product strategy — focused on enterprise and safety — can beat a consumer-scale strategy. That's the opposite of what the market believed in 2024.
For OpenAI. Spud needs to deliver. Not just a good model — a model that reassures the CIOs who are hesitating. If GPT-6 disappoints like GPT-4.5 did, the gap with Anthropic could double by the end of 2026.
TL;DR:
- Anthropic overtakes OpenAI for the first time since 2022 with $30B ARR versus $25B for OpenAI (April 2026)
- Growth: $1B January 2025 → $9B December → $14B February → $30B April 2026 = 30x in 15 months
- 1,000+ enterprise customers spending over $1M annually, doubled in less than 2 months after Series G
- Anthropic spends 4x less than OpenAI to train its models (~$5B vs ~$20B)
- IPO targeted for October 2026 at a $380B valuation to raise $60B — the largest tech IPO since Aramco
Anthropic passes OpenAI. Not in communications, not in valuation, not in consumer user count — in real, signed, recurring revenue. The winning model isn't the one that tweets loudest, it's the one CIOs put into production. Dario Amodei built a company that burns four times less compute to generate 20% more revenue. If capital-efficiency logic ultimately takes hold in AI the way it did in cloud between 2015 and 2020, Anthropic isn't just passing OpenAI in April 2026. It's redefining what a viable AI company looks like.
Sources: Trending Topics — Anthropic $30B run rate, SaaStr — 4x less compute, Whalesbook — revenue surges, TradingKey — IPO October 2026.
