Anthropic revenue hits $30B ARR after tripling in four months, surpassing OpenAI's $24B as over 1,000 enterprises pay $1M+ annually and eight of the Fortune 10 adopt Claude.

4-Month Triple to $30B ARR: Anthropic Revenue Tops OpenAI

by
💡 Key Takeaways

  • Anthropic revenue hit a $30B ARR in April 2026, roughly 3.3x its $9B exit run-rate from year-end 2025 — and ahead of OpenAI’s $24B.
  • Enterprise clients paying $1M or more per year doubled from 500 to 1,000+ in two months, with eight of the Fortune 10 now running on Claude.
  • The gap is narrower than the headline suggests: Anthropic books gross revenue (partner cuts counted as costs), while OpenAI reports net receipts after cloud-share payouts.

$30 billion. Four months ago, that number belonged to OpenAI alone. On April 7, 2026, Anthropic revenue on an annualized run-rate basis (ARR, a standard SaaS gauge that multiplies monthly revenue by twelve) crossed $30B — and the company became the first AI startup to pass the chief rival that dominated the sector for three straight years.

The speed matters more than the headline. Anthropic revenue tripled in roughly one quarter, jumping from $9B at year-end 2025. OpenAI, by contrast, climbed from $20B to $24B over the same window — a respectable 1.2x that still lost the footrace.

This piece unpacks the $30B story from five angles: the growth curve, the accounting asterisk that narrows the gap, the enterprise strategy that flipped the consumer-led AI hierarchy, the Claude Code flywheel behind the developer pull, and the compute and valuation tells investors should track next.

Anthropic revenue $30B ARR surpasses OpenAI $24B AI company growth illustration (AI-generated)
Anthropic revenue topping $30B reframes the AI industry’s revenue hierarchy — Generated by OpenAI DALL-E 3 (AI-generated)

Anthropic Revenue Hits $30B — A 4-Month Triple That Broke the AI Pecking Order

Sherwood News first reported on April 8, 2026 that Anthropic’s ARR had topped $30B after a Monday disclosure. That is approximately 3.3x the $9B ARR the company exited 2025 with. In practical terms, Anthropic added a full OpenAI-2024 ($9B) plus a Box-sized SaaS business on top — in roughly 120 days.

Stretch the curve longer and the acceleration becomes harder to square with normal SaaS economics. Per SaaStr’s tally, Anthropic’s ARR went from $87M in January 2024 to $1B by December 2024, $9B by December 2025, and past $30B this April. Meritech’s Alex Clayton, a partner who has studied more than 200 public software IPOs, said he has “never saw a growth rate like this.”

Even professional forecasters whiffed. Epoch AI had predicted in February 2026 that Anthropic might pass OpenAI around mid-year, a view echoed by analyst desks that priced in the crossover as a 2H 2026 event. The actual overtaking arrived more than two months early, compressing the window for OpenAI to defend its lead with product launches or pricing moves.

The revenue curve also resets how public-market investors should think about AI-sector comparables. A business that tripled ARR in one quarter belongs in its own bucket, separate from the 40-50% growth rates that used to define elite SaaS. For allocators underwriting AI infrastructure plays, the question is no longer whether Anthropic is a serious commercial outfit — it is whether any of the historical SaaS valuation multiples still apply at this pace.

OpenAI’s $24B vs. Anthropic’s $30B — The Accounting Asterisk You Can’t Ignore

OpenAI disclosed ARR of about $24B at the end of March 2026, implying roughly $2B a month in revenue. That was up from $20B at year-end 2025, a 1.2x run — strong in absolute terms, but visibly slower than Anthropic’s triple-digit-percent quarter.

OpenAI CFO Sarah Friar framed the growth in superlative terms, telling staff and press the company was seeing “never-before-seen growth at such scale.” The quote matters because it undercuts any “OpenAI is losing” narrative. The more accurate reading: both labs are growing in a way the software industry has never witnessed, and one is simply accelerating faster than the other.

Anthropic $30B ARR vs OpenAI $24B ARR revenue comparison on mobile trading screen
Headline ARR hides accounting mechanics — Anthropic’s gross basis lifts the reported number well above net receipts — Photo by Jakub Żerdzicki on Unsplash

Sherwood warned bluntly that the comparison “isn’t perfectly apples to apples.” Anthropic recognizes the full amount its cloud customers pay as revenue, treating partner rake as a cost of goods sold (a gross-basis convention, where the seller records the full buyer payment and deducts partner share as an expense). OpenAI, on the other hand, books only what it actually receives after cloud partners take their slice (a net-basis convention).

The Information estimates Anthropic will pay around $1.9B to cloud partners in 2026 and $6.4B in 2027. Strip those payouts from the top line and the real revenue gap is materially tighter than the $6B headline difference — and could flip again if Anthropic’s partner-share obligations scale faster than OpenAI’s net-basis receipts.

Key point: Anthropic revenue leads on the reported figure, but gross-versus-net accounting compresses the real gap — and Anthropic’s projected $6.4B cloud-partner payout in 2027 could reverse the rank order on a like-for-like net basis.

1,000 Clients Paying $1M+ — Why Eight of the Fortune 10 Picked Claude

The growth engine is unambiguous: large enterprises. Sherwood reported that the number of customers spending over $1M a year with Anthropic went from roughly 500 in February 2026 to more than 1,000 by April — a doubling in two months. That is not viral consumer growth; it is procurement-committee growth, which is far harder to pull off and more durable when it does happen.

Fortune 10 enterprise office running Claude for internal AI workloads
Eight of the Fortune 10 are now Anthropic customers, anchoring roughly 80% of revenue — Photo by blue sky on Unsplash

SaaStr’s analysis notes that eight of the Fortune 10 are already Claude customers, and that Anthropic ended October 2025 with more than 300,000 business accounts contributing roughly 80% of company revenue. Put simply, this is a B2B franchise masquerading as an AI lab — and the Stanford AI Index’s 1,503 Elo score (covered in our breakdown of the benchmark reshuffle) now doubles as procurement justification for CIOs.

The strategic contrast with OpenAI is sharp. Anthropic focused on API integration and embedded enterprise products from day one, while OpenAI’s flywheel has been ChatGPT subscriptions plus a growing consumer-AI portfolio (see our analysis of the GPT-5.5 Spud pivot). OpenAI still counts enterprise at 40%+ of revenue and has 9 million paid business users, but when you weight by marquee-logo concentration, the tilt reverses.

Put simply, the two companies were never really fighting for the same customer. Anthropic’s revenue base skews toward a small number of very large contracts, which maximizes annual contract value but exposes the company to concentration risk if even a few Fortune 10 customers renegotiate. OpenAI, with nine million paid business seats, captures the long tail — a broader but thinner base that is harder for competitors to dislodge one seat at a time.

Claude Code Is the Wedge — 4% of Public GitHub Commits Come From One Tool

If there is a single product pulling enterprise dollars through the Anthropic funnel, it is Claude Code. Per SaaStr, the AI coding tool hit a $2.5B annualized run-rate by February 2026. In other words, one product inside Anthropic matches the entire 2024 revenue of most publicly traded AI SaaS firms.

SaaStr also reports that Claude Code now authors roughly 4% of public GitHub commits — a footprint that would take most developer tools a decade to build. OpenAI’s Codex is not far behind: it counts 2 million weekly users and is growing at a 70% month-over-month pace. The takeaway: the coding-agent market has become the decisive front of the AI revenue war.

Why this matters: developer adoption creates a bottom-up pull that feeds top-down procurement. An engineer who uses Claude Code at home ships better work at the office, advocates internally, and within a few quarters her company is signing a seven-figure enterprise contract. Put simply, the same seat that used to be a $20-per-month GitHub Copilot sale has become the wedge for $1M+ Anthropic deals.

In simpler terms: Claude Code works like a Trojan horse for enterprise sales. Individual developers bring it inside the firewall, their teams standardize on it, and procurement eventually writes a million-dollar check to legitimize what the engineering org is already doing.

Compute Race — $30B Training Budget Meets a Fresh 3.5GW Pact

Revenue without compute is a ceiling. Per SaaStr, Anthropic’s peak training bill runs near $30B, while OpenAI’s planned training spend sits roughly 4x higher. The implication is uncomfortable for OpenAI: Anthropic is outgrowing its rival on revenue while spending materially less to build the underlying models.

Anthropic Google Broadcom 3.5GW AI compute supply data center server racks
Anthropic’s Google and Broadcom deal locks in 3.5GW of incremental AI compute beginning in 2027 — Photo by Domaintechnik Ledl.net on Unsplash

The OpenAI side of the ledger looks even heavier. SaaStr cites projections showing OpenAI plans $121B in compute spend and $85B of losses in 2028 alone. Friar’s growth quote notwithstanding, the company is building a P&L that will need capital-markets patience on an almost unprecedented scale. Strip training costs from both models and each lab is close to operating profitability — which is a reminder that today’s losses are mostly a bet on the next generation of models.

On the supply side, Trendingtopics.eu notes Anthropic locked in 3.5GW of Google and Broadcom compute that begins coming online in 2027. The key here is sequencing: if ARR is to keep tripling, the compute supply curve has to stay in front of demand. A 3.5GW pre-commitment signals Anthropic expects Claude API demand to keep compounding through the back half of the decade.

Key point: Anthropic leads on revenue while spending roughly a quarter of OpenAI’s training budget, making capital efficiency the new scoreboard — and the 3.5GW Google and Broadcom deal funds the next leg of ARR growth ahead of the 2027 compute crunch.

Valuation and Capital — $380B vs. $852B, and What Investors Should Watch

Revenue parity has not translated into valuation parity. Per Trendingtopics.eu, Anthropic was valued at $380B as of February 2026, with total funding now past $18B and Forbes pegging CEO Dario Amodei’s personal net worth at roughly $7B. That is a remarkable outcome for a company that passed its third birthday this year.

OpenAI remains the heavier weight on the capital side. It closed a $122B financing at an $852B valuation, per SaaStr. On an ARR-to-valuation basis Anthropic trades at roughly 7.9% — more than 2x the 2.8% implied ratio for OpenAI — a spread that could tighten quickly if growth continues to favor the smaller lab.

Two numbers should anchor every investor’s watchlist from here. (1) Whether OpenAI’s projected 2028 compute spend of $121B and $85B loss print continue to hold as quarterly data comes in, and (2) whether Anthropic’s estimated $6.4B cloud-partner payout in 2027 actually materializes and flips the net-basis ranking back. Those are the two tells that will decide the next chapter of the AI revenue race.

💡 Related Posts
1,503 Elo: Anthropic Dethrones OpenAI After 3-Year Reign
💡 Related Posts
OpenAI Axed Sora and Bet It All on GPT-5.5 ‘Spud’

References

Stay Ahead of the Curve

Get the latest insights on AI, tech, and markets — delivered in both Korean and English.

Explore More Posts