AI Revenue (ARR) Leaders
The ten AI-native companies with the highest reported annualized revenue run rates, as of the latest public figures. Every number carries a confidence tier; an external link appears only when it meets the sourcing policy, because private revenue is a claim, not a filing. Figures are re-checked as new reports land.
| # | Company | Value | Date | Tier |
|---|---|---|---|---|
| 1 | Anthropic Bloomberg-reported run rate ahead of a planned IPO — up from ~$9B a year earlier. | $65B+ | 2026-07-31 | Reported |
| 2 | OpenAI Roughly $2B a month of revenue against a compute program far larger. | ~$25B | 2026-02-28 | Reported |
| 3 | Databricks Company-disclosed run rate at the $5B round — data-and-AI platform, not a lab. | $6.9B | 2026-06-30 | Disclosed |
| 4 | Cursor (Anysphere) Fastest to $100M ARR in its cohort (12 months) — acquired by SpaceX weeks after this print. | $4B+ | 2026-06-30 | Reported |
| 5 | Crusoe Neocloud revenue — energy-first AI infrastructure; on track for ~US$2B revenue this year per Newcomer. | ~$2B | 2026-09-23 | Reported |
| 6 | Mercor AI recruiting and expert-data marketplace, The Information-reported. No approved external source URL is currently attached. | ~$2B | 2026-06-30 | Reported |
| 7 | Zhipu AI Reported via 36Kr and listing disclosures — the only listed company on the board. No approved external source URL is currently attached. | $1.6B | 2026-08-31 | Reported |
| 8 | Lambda GPU-cloud run rate, research-reported. No approved external source URL is currently attached. | ~$1.5B | 2026-08-31 | Reported |
| 9 | Surge AI Data-labeling rival to Scale, Reuters/Sacra-reported; figure is the oldest on the board. No approved external source URL is currently attached. | $1.4B | 2025-08-31 | Reported |
| 10 | Scale AI Post-Meta-deal revenue, with Google and OpenAI among lost customers. No approved external source URL is currently attached. | $1B+ | 2026-08-31 | Reported |
ARR definitions differ by company: some quote annualized monthly revenue, others net revenue after credits or reseller splits. Foundation-model labs dominate the top of the board; only two companies below the $1B line make the cut. Pre-AI transformed SaaS businesses are excluded by definition.
More boards
Biggest AI Acquisitions & Control Deals
#1 · xAI ($250B)
Highest AI Company Valuations
#1 · Anthropic ($965B)
Largest AI Funding Rounds
#1 · OpenAI ($122B)
The full capital picture lives on the Funding Leaderboard.
Frequently asked questions
Why does Anthropic's $65B run rate dwarf OpenAI's $25B when both are frontier labs?
Timing and definition. Anthropic's figure is a July 2026 Bloomberg-reported run rate after a year of steep enterprise growth; OpenAI's $25 billion is a February 2026 report, four months stale. OpenAI's own run rate has likely grown since, but no newer whitelisted figure exists — the board shows the latest verifiable number for each company, not a synchronized snapshot.
What counts as AI-native revenue on this board?
Revenue from AI products themselves — model access, AI applications, AI compute and data services. A traditional SaaS company bolting AI features onto an existing product does not qualify, which is why several larger software franchises are absent: the board prices the AI economy, not the software economy that added AI.
Why is Databricks here when it started as a data platform?
Because its revenue now runs through AI workloads — model serving, vector data and MosaicML-derived training revenue — and the company itself discloses the $6.9 billion figure. The board's criterion is what generates the revenue today, not what the company sold at founding; Databricks disclosed its number, which is why it carries the Disclosed tier.
How reliable are run-rate figures compared with real revenue?
A run rate is monthly revenue annualized — it assumes the month represents the year. For fast-growing companies it understates, for decelerating ones it overstates, and definitions vary: some quote gross revenue before inference-credit costs, others net. Every row carries a tier; source links appear only when they meet the published sourcing policy.
Why does the board thin out below $1.5B?
Because private AI companies report revenue selectively — usually to support a raise or a narrative. Below roughly $1 billion, figures arrive as research estimates with wide error bars, and the sourcing rule prefers fewer verified rows over a fuller board of guesses. Scale AI's $1B+ and Fireworks' disclosed $1B+ sit at the cut line.
Which company grew fastest to its current ARR?
Cursor. Anysphere reported reaching $100 million of ARR within twelve months of launch — among the fastest revenue ramps ever recorded in software — and roughly $4 billion of run rate within two years, before the SpaceX acquisition. Lovable's $1M-to-$500M curve is the compounding champion, but it started from consumer pricing rather than enterprise seats.