Is Anthropic now ahead of OpenAI?
Every funding round, every investor, and what the labs promised in return, in one document. Have a look ↓
Summary
Yes, on revenue. Anthropic is now ahead of OpenAI on every recent reported revenue measure, even after adjusting for the accounting gap between them. It is not ahead on users. To answer “Is Anthropic now ahead of OpenAI?”, we analyzed company-reported run-rates, reported quarterly revenue, ChatGPT and Claude usage, Claude Code adoption, enterprise signals, valuations, and compute commitments.
The crossover happened in Q2 2026. According to Reuters and the Wall Street Journal, Anthropic booked about $11.6B that quarter, versus $6.7B for OpenAI. In Q1 2026, OpenAI was still ahead, at $5.7B versus $4.7B.
The lead held through the latest numbers. Bloomberg reported Anthropic above $65B in annualized revenue at the end of July 2026. OpenAI told investors it was near $50B at the end of September 2026. In October 2026, Anthropic told investors its Q3 revenue was $16.8B.
The accounting caveat matters, but it does not flip the answer. Anthropic books cloud-partner sales gross; OpenAI books some of them net. A source quoted by Axios said net reporting would cut Anthropic’s revenue by about 6% to 10%. That still leaves Anthropic ahead.
Anthropic’s growth is slowing. Quarterly revenue grew about 147% from Q1 to Q2 2026, then about 45% from Q2 to Q3. That is still fast, but it is the main question hanging over its planned November IPO.
This was not a mass exodus from ChatGPT. OpenAI reported more than 1.2B weekly ChatGPT users in October 2026. Anthropic still does not disclose Claude user numbers.
Claude is taking share on the web, though. Similarweb estimates Claude’s share of visits to major AI chatbot sites rose from 2% to 9.6% in the 12 months to August 2026, while ChatGPT’s fell from 78% to 57%.
Anthropic’s edge is professional use: enterprise accounts, API workloads, cloud channels, and coding. JetBrains found 39% of professional developers used Claude Code at work in May–July 2026, versus 16% for Codex.
On valuation, Anthropic’s last priced round ($965B, May 2026) is above OpenAI’s ($852B, March 2026). The ~$2T discussed for Anthropic’s IPO and the $1.4T OpenAI is reportedly seeking are targets, not prices.
The sharpest conclusion: ChatGPT still wins on reach, and Claude now wins on revenue. But Anthropic also carries over $600B in reported noncancelable compute commitments, so being ahead on sales does not yet mean being the safer business.
Written by the New Market Pitch teamWe track new markets so founders and investors can move faster. About us
We keep an up-to-date report on frontier AI labs. Have a look ↓
Sample pages from an earlier edition. The report has been updated with new data since.
Can you compare OpenAI and Anthropic revenue like for like?
Not exactly, but closely enough to see who leads. Revenue is still the right metric. The catch is that the two companies count it differently.
Users are not a good alternative. OpenAI reports ChatGPT weekly users: 900M+ in March 2026, 1B+ on August 31, 2026, and 1.2B+ in October 2026. Anthropic reports no Claude-wide user number at all. Third-party traffic data exists, but Similarweb itself notes it excludes API usage, which is where much of Anthropic’s business sits.
Paying users are messy too. A ChatGPT Plus subscriber, a Claude Code seat, an API customer, and a seven-figure enterprise contract are all “paid usage,” but they are very different businesses.
So revenue it is. Since mid-2026, we have three kinds of figures: official run-rate checkpoints, run-rates reported by the press, and quarterly revenue reported from investor and IPO documents. None of the 2026 quarterly figures is audited yet.
The bigger problem is accounting. When a customer buys Claude through AWS or Google Cloud, Anthropic books the full sale as revenue and the cloud provider’s cut as a cost. OpenAI books only its own share of some partner sales, such as those through Microsoft. Bloomberg reported this difference on October 9, 2026. Reports tied it to the gap between an earlier reported OpenAI run-rate near $70B and the roughly $50B OpenAI gave investors.
How big is the distortion for Anthropic? A source familiar with its financials told Axios in September 2026 that net reporting would cut its revenue by about 6% to 10%. That is meaningful, but not enough to erase a lead of 30% or more. The full answer should come with Anthropic’s public prospectus.
If you want more recent data on this point, see the latest figures in our Frontier AI Labs report ↓.
What have OpenAI and Anthropic said about revenue in 2026?
Both companies have shared far more in 2026 than before, mostly with investors ahead of possible IPOs. Most of the newest numbers come through press reports, not official releases.
OpenAI’s official series is $2B ARR in 2023, $6B in 2024, and $20B+ in 2025. In March 2026, it said it was earning about $2B a month. Since then, the figures are reported. Quarterly revenue was $5.7B in Q1 and, according to Reuters, about $6.7B in Q2. The annualized run-rate is a different measure: it takes the latest month and multiplies it by 12. It was $40B+ in August, according to Bloomberg, and almost $50B in September, OpenAI told investors, according to Reuters and the FT. That implies very fast growth since Q2, and it rests on reported, unaudited figures. OpenAI also has a reported target of about $70B in annualized revenue by December 2026.
Anthropic’s official series is roughly $1B run-rate in January 2025, $5B+ in August 2025, $14B in February 2026, $30B+ in April 2026, and $47B+ in May 2026. Reported figures since: $4.7B in Q1 revenue, $11.6B in Q2, $65B+ annualized at the end of July, and $16.8B in Q3.
The pattern is clear. OpenAI has the longer record. Anthropic has had the higher current revenue since Q2 2026.
| Date | Company | What was communicated | Comment |
|---|---|---|---|
| October 2026 | Anthropic | $16.8B Q3 2026 revenue (reported) | Shared with investors on October 14, 2026, ahead of a planned November IPO. Up about 45% from Q2. Reports also cite an 11% adjusted operating margin and positive adjusted free cash flow. None of this is audited. The firm SemiAnalysis estimated the current run-rate at about $74B; other analysts expect about $110B by year-end. Both are estimates. |
| September 2026 | OpenAI | ~$50B annualized run-rate (reported) | First reported by the Financial Times and confirmed by Reuters and Bloomberg on October 8, 2026. It replaced an earlier report of nearly $70B; the gap was reportedly tied to how cloud-partner sales are counted. OpenAI’s actual 2026 revenue is projected at about $35B to $36B, according to Bloomberg and the FT. |
| August 2026 | OpenAI | $40B+ annualized run-rate (reported) | Shared internally by co-founder Greg Brockman, according to Bloomberg. On August 31, 2026, OpenAI also said ChatGPT Ads had reached a $1B run-rate in under 200 days. |
| End of July 2026 | Anthropic | $65B+ annualized run-rate (reported) | Reported by Bloomberg on August 17, 2026. Not confirmed by Anthropic. Even with a 6% to 10% net-accounting cut, it stays above OpenAI’s ~$50B in September. |
| Q2 2026 | Both | Anthropic ~$11.6B vs OpenAI $6.7B (reported) | The only same-quarter comparison. Anthropic more than doubled from Q1; OpenAI grew about 18%, per the WSJ. Bloomberg put Anthropic at $11.5B; Reuters at $11.6B. The small gap does not change the read. |
| May 2026 | Anthropic | $47B+ revenue run-rate | Anthropic’s latest official run-rate, announced with its $65B Series H at a $965B post-money valuation on May 28, 2026. |
| April 2026 | Anthropic | $30B+ revenue run-rate | Up from roughly $9B at the end of 2025. The first official figure above OpenAI’s last official numbers ($20B+ for 2025, about $24B a year in March 2026). |
| March 2026 | OpenAI | ~$2B monthly revenue | About $24B a year. OpenAI also said enterprise was more than 40% of revenue. This is its last official revenue figure. |
| February 2026 | Anthropic | $14B revenue run-rate | 500+ customers spent over $1M a year, and Claude Code was above $2.5B. Anthropic was still smaller than OpenAI. |
| January 2026 | OpenAI | $20B+ ARR for 2025 | OpenAI’s CFO gave the series $2B (2023) → $6B (2024) → $20B+ (2025), and linked growth to available compute. |
| End of 2025 | Anthropic | ~$9B revenue run-rate | Above OpenAI’s 2024 level, but far below its 2025 level. Anthropic’s actual 2025 revenue was nearly $4.6B, per its draft prospectus as reported by Reuters. |
| August 2025 | Anthropic | $5B+ revenue run-rate | Up from about $1B at the start of 2025. Claude Code passed $500M soon after, and $1B in November 2025. |
| 2024 | OpenAI | $6B ARR | OpenAI tripled from 2023. It was about 6x larger than Anthropic’s early-2025 baseline. |
| Beginning of 2025 | Anthropic | ~$1B revenue run-rate | The base for the later ramp: $65B+ was reported 19 months later. |
| 2023 | OpenAI | $2B ARR | The first point in OpenAI’s series, after ChatGPT had already become a mass-market product. |
These pages are excerpts from our report. See the rest ↓
Sample pages from an earlier edition. The report has been updated with new data since.
When did Anthropic overtake OpenAI on revenue?
In the second quarter of 2026. Official run-rates showed the flip first, in April 2026. Reported quarterly revenue confirmed it in Q2: about $11.6B for Anthropic versus $6.7B for OpenAI.
OpenAI led for a long time, and by a wide margin. It launched ChatGPT in November 2022 and reached $2B ARR in 2023. Anthropic launched Claude publicly in 2023 and reported actual revenue of only about $386M in 2024. Around early 2025, OpenAI was about 6x larger.
The gap shrank in 2025. By year-end, OpenAI was at $20B+ ARR and Anthropic at about $9B. OpenAI was still more than 2x larger. In Q1 2026, OpenAI still led on actual revenue, $5.7B to $4.7B.
Then the curves crossed. Anthropic’s revenue more than doubled in Q2 2026, while OpenAI’s grew about 18%. Since then, every reported figure has kept Anthropic ahead: $65B+ at the end of July versus OpenAI’s ~$50B at the end of September, and $16.8B in Q3 revenue. OpenAI has no reported Q3 number yet.
Does the accounting change the date? Probably not. Cut Anthropic’s Q2 revenue by 10% for net reporting and it is still about $10.4B, well above OpenAI’s $6.7B. The accounting gap changes the size of the lead, not its existence.
If you want more recent data on this point, see the latest figures in our Frontier AI Labs report ↓.
| Period | OpenAI revenue scale | Anthropic revenue scale | Gap / read |
|---|---|---|---|
| 2023 | $2B ARR | No clean disclosed number | OpenAI clearly ahead. |
| 2024 → early 2025 | $6B ARR | ~$1B run-rate | OpenAI about 6x larger. |
| End of 2025 | $20B+ ARR | ~$9B run-rate | OpenAI still more than 2x larger. |
| Q1 2026 | $5.7B revenue (reported) | $4.7B revenue (reported) | OpenAI still ahead, by about 1.2x. |
| Q2 2026 | $6.7B revenue (reported) | ~$11.6B revenue (reported) | Crossover: Anthropic about 1.7x larger before any net adjustment. |
| July–September 2026 | $40B+ (Aug.), ~$50B (Sep.), reported | $65B+ (end of July), reported | Anthropic still ahead, even after a 6% to 10% net cut. |
| Q3 2026 | Not reported | $16.8B revenue (reported) | No same-quarter comparison yet. |
Is Anthropic’s revenue growth slowing down?
Yes. Percentage growth fell sharply, and the dollars added per quarter fell too, though less.
Anthropic’s quarterly revenue went from $4.7B in Q1 2026 to $11.6B in Q2, a gain of about 147%. Q3 reached $16.8B, a gain of about 45%. The amount added per quarter fell from about $6.9B to about $5.2B.
That matters because Anthropic is going public on its growth story. Some investors had expected a year-end run-rate above $100B. Analysts who met management in October 2026 put it near $110B; SemiAnalysis estimated the current pace at about $74B. These are estimates, not company disclosures, and the gap between them shows how much Q4 has to deliver.
Slower growth was always likely. A company cannot keep doubling every quarter at a $60B+ scale. The real test is whether Q4 adds more than Q3 did. If quarterly gains keep shrinking, a ~$2T IPO valuation gets harder to defend.
OpenAI is not the one gaining here, though. Its reported quarterly growth was about 18% in Q2 2026. On percentage growth, Anthropic is still much faster.
Prefer verified numbers to headlines? Our report is here ↓
Sample pages from an earlier edition. The report has been updated with new data since.
Is Anthropic worth more than OpenAI now?
On the last priced rounds, yes. On the targets now being discussed, it depends on two deals that have not happened yet.
Anthropic raised $65B at a $965B post-money valuation on May 28, 2026. OpenAI’s last completed round was $122B at $852B on March 31, 2026. That puts Anthropic ahead on the most recent actual prices.
Both companies want more. Anthropic is reportedly preparing an IPO for mid-November 2026, after the US midterm elections, with Goldman Sachs and Morgan Stanley reportedly leading. Around $2T has been discussed; some investors estimate closer to $1.5T to $1.8T. OpenAI deferred its IPO to 2027 at the earliest and is reportedly seeking at least $30B at a $1.4T valuation. Neither number has been achieved.
On revenue multiples, the gap narrows. At $965B, Anthropic was priced at about 20x its $47B+ run-rate in May, and below 15x the $65B+ reported for July. At its proposed $1.4T, OpenAI would be about 28x its ~$50B run-rate. A $2T Anthropic IPO would be about 30x its annualized Q3 revenue. Investors are paying a similar premium for both.
Being bigger also means bigger obligations. Anthropic’s draft prospectus reportedly lists over $600B in noncancelable infrastructure commitments, per the WSJ in October 2026, and an $80.7B net loss for H1 2026, mostly from financing-related accounting. Reuters also reported that Amazon and Google routed 47% of Anthropic’s 2025 sales, and that two unnamed customers each made up 12%. Its 11% adjusted operating margin in Q3 is non-GAAP and does not cover those commitments.
If you want more recent data on this point, see the latest figures in our Frontier AI Labs report ↓.
Why is Anthropic beating OpenAI on revenue?
Anthropic did not become “the new ChatGPT.” It won because it sold to professionals who pay much more per user than consumers do.
OpenAI was not falling apart. It went from $2B ARR in 2023 to a reported ~$50B run-rate in September 2026. That is strong growth. Anthropic’s curve was simply steeper: roughly $1B in January 2025 to a reported $65B+ by the end of July 2026.
The first reason is coding. Claude Code launched publicly in May 2025, passed $1B run-rate in November 2025, and $2.5B+ by February 2026. Anthropic has not given a newer official figure. Independent data shows the lead held: JetBrains surveyed more than 15,000 developers and found 39% used Claude Code at work in May–July 2026 (47% in the US), versus 16% for Codex. Codex grew about fivefold from January, so OpenAI is catching up from a low base.
Coding is one of the easiest AI budgets to justify. A consumer pays $20 a month. An engineering team can spend six or seven figures a year if the tool saves hours. In October 2026, Barclays said it aims to have half its developers on Claude Code by the end of 2026.
The second reason is big enterprise accounts. Anthropic had 300,000+ business customers by September 2025, and 1,000+ customers spending $1M+ a year by April 2026, double the February count. Menlo Ventures estimated Anthropic took 40% of enterprise LLM API spending in 2025, versus 27% for OpenAI, and 54% of coding API spend. No newer survey was found.
The third reason is distribution through the clouds. Claude is sold on AWS Bedrock, Google Cloud Vertex AI, and Microsoft Azure Foundry. Over 100,000 customers ran Claude on Bedrock by April 2026. Buyers could use existing cloud budgets and security reviews. The flip side is dependence: Amazon and Google routed nearly half of Anthropic’s 2025 sales.
The fourth reason is compute and models. Anthropic secured multi-gigawatt capacity from Google, AWS (over $100B over 10 years), and Microsoft, which let it serve demand. Its models stayed at the top: on October 19–20, 2026, Claude Opus 5.5 ranked first on Arena’s web-development and agent leaderboards and first on the Artificial Analysis Intelligence Index, ahead of OpenAI’s GPT-6 Astra.
The common thread is budget reallocation. Companies moved money from developer tools, contractors, and software seats into Claude usage. OpenAI is responding with Codex, enterprise deals such as Atlassian, and ads. But on the highest-paying workloads, Anthropic got there first.
| Driver | What happened / why it mattered |
|---|---|
| Enterprise accounts | $1M+ customers doubled to 1,000+ by April 2026; an estimated 40% of 2025 enterprise LLM API spend. |
| Claude Code | $2.5B+ run-rate by Feb. 2026; used at work by 39% of developers in mid-2026, versus 16% for Codex. |
| Cloud distribution | AWS, Google Cloud, and Azure made buying easy; Amazon and Google routed 47% of 2025 sales. |
| High-ROI workflows | Coding and API workloads justify large budgets because they save expensive engineering time. |
| Compute and models | Multi-gigawatt deals let Anthropic serve demand; Opus 5.5 led coding and agent leaderboards in October 2026. |
If you want more recent data on this point, see the latest figures in our Frontier AI Labs report ↓.
Want to know who is really ahead? Our report is here ↓
Sample pages from an earlier edition. The report has been updated with new data since.
Did people abandon ChatGPT for Claude?
No. ChatGPT is still growing fast. But Claude has become a real second choice, and it is taking share.
ChatGPT went from 900M+ weekly users in March 2026 to 1B+ at the end of August and 1.2B+ in October 2026, when OpenAI rolled out GPT-6 in ChatGPT. A product people are leaving does not add 300M weekly users in seven months.
Claude’s gains are real, though. Similarweb estimates Claude’s share of visits to major AI chatbot sites rose from 2% to 9.6% in the 12 months to August 2026, while ChatGPT’s fell from 78% to 57%. In the same period, the share of ChatGPT users who also used Claude rose from 2% to 14%. Claude’s six-month retention improved from about 26% to 45%, though ChatGPT’s is still higher at 59%.
What does that mean? People are adding Claude, not leaving ChatGPT. Many now use both. The chatbot market grew so fast that ChatGPT could lose share and still gain users.
The revenue gap closed mainly through professional spending, not consumers. 1,000+ customers spending $1M+ a year matter more to Anthropic’s revenue than tens of millions of free users.
The distinction is simple: ChatGPT has more people; Claude has more revenue in high-value workflows. A casual user may pay nothing. A Plus user pays $20 a month. A company using Claude for engineering can pay six or seven figures a year.
Do more people use ChatGPT or Claude today?
ChatGPT, by a very wide margin.
The cleanest number is OpenAI’s own: more than 1.2B weekly ChatGPT users in October 2026. Anthropic still publishes no Claude-wide user count, and outside estimates vary widely. The closest comparison is Sensor Tower’s app data: about 1B monthly active app users for ChatGPT in Q2 2026, versus about 56M for Claude, roughly 18x more.
Web traffic shows a smaller gap. Similarweb’s August 2026 estimate gives ChatGPT 57% of visits to major AI chatbot sites, versus 9.6% for Claude, about 6x more. Claude’s sign-up pages drew 21.4M visits that month, up 504% year over year.
The two gaps differ because they measure different things. A likely reason is that Claude is used more on desktop and at work, which web data captures better than app data. Neither measure is a company-reported user count.
Claude is growing much faster in percentage terms. But from a small base, high growth still leaves it far behind a product with over a billion weekly users.
The comparison in a table:
| Metric | ChatGPT | Claude | Read |
|---|---|---|---|
| Company-reported weekly users | 1.2B+ WAU (Oct. 2026) | Not disclosed | OpenAI gives a clean usage number; Anthropic does not. |
| Third-party monthly app users | ~1B MAU (Q2 2026) | ~56M MAU (Q2 2026) | ChatGPT is roughly 18x larger on app usage. |
| Share of AI chatbot web visits | 57% (Aug. 2026), from 78% | 9.6% (Aug. 2026), from 2% | Claude is gaining share, from a small base. |
| Six-month retention (web cohorts) | 59% | 45% (Feb. 2026 cohort) | Claude users are sticking around more, but ChatGPT still keeps more. |
| Revenue story | Huge consumer base + paid tiers + ads + API + enterprise | Smaller user base, denser professional monetization | Claude leads on reported revenue without matching ChatGPT usage. |
Revenue leadership and user leadership are not the same thing.
Anthropic moved ahead on revenue in 2026 without coming close to ChatGPT on users.
It did it by monetizing a much smaller base more intensely, through professional, coding, API, and enterprise use.
ChatGPT is still the default mass-market AI product. Claude is a much smaller product with unusually high revenue density.
If you want more recent data on this point, see the latest figures in our Frontier AI Labs report ↓.
We made a report out of all this. Take a look ↓
Sample pages from an earlier edition. The report has been updated with new data since.
Does Claude make more money per user than ChatGPT?
Yes, by a very wide margin, even if no exact figure exists.
Neither company publishes revenue per user, so we can only make a rough calculation. Take OpenAI’s reported ~$50B run-rate and ChatGPT’s ~1B monthly app users: about $50 per user per year. Take Anthropic’s reported $65B+ and Claude’s ~56M app users: about $1,160 per user per year. That is roughly 23x more.
The real gap is smaller than that. Both revenue figures include API, enterprise, and cloud sales, not only app users. Anthropic’s figure is partly gross, and app data misses much of Claude’s desktop and work use. Even if you cut Claude’s ratio in half, it stays more than 10x higher.
That is the core of the story. Anthropic’s revenue does not depend on its consumer app. It comes from fewer, more valuable users: developers, enterprise teams, and companies running Claude through the cloud.
OpenAI is working to raise its revenue per user. ChatGPT Ads reached a $1B run-rate in under 200 days, OpenAI said on August 31, 2026. It also had 9M+ paying business seats in March 2026 and expects enterprise to reach about half its revenue by the end of 2026. With 1.2B weekly users, even a small gain per user adds up to a lot of money.
So yes, Anthropic earns much more per user, but precisely: Claude does not earn more from the average consumer; Anthropic earns more because its users are mostly professionals. ChatGPT wins on reach. Claude wins on revenue density.
If this is useful, the full report goes much further. Have a look ↓
Sample pages from an earlier edition. The report has been updated with new data since.
Our methodology
This analysis tests whether Anthropic is now ahead of OpenAI based on the best public evidence available as of October 2026: company-disclosed ARR, revenue run-rate checkpoints, quarterly revenue reported by the press from investor and IPO documents, usage disclosures, Claude Code monetization, enterprise customer counts, cloud distribution, compute capacity, and third-party usage estimates.
We did not fill missing years with speculative revenue estimates. When a company had not disclosed a clean comparable number, we left the gap visible and used cautious wording rather than inventing a figure.
We compared the companies through the best available checkpoints, even when the periods were not perfectly matched. Some numbers are annual ARR, some are point-in-time run-rates, and some are quarterly revenue reported from unaudited documents. We labeled reported figures as such and did not treat them as audited results.
We used revenue as the core comparison because user metrics are not disclosed symmetrically. OpenAI reports broad ChatGPT usage, while Anthropic shares narrower signals for Claude Code and enterprise adoption, so a direct user-to-user comparison would mix unlike products and unlike data sources.
We treated Anthropic’s reported Q2 2026 revenue lead and its later run-rates as evidence that it is ahead on revenue. Because Anthropic books cloud-partner sales gross and OpenAI books some of them net, we tested the lead against the 6% to 10% net adjustment reported in the press and did not state a precise percentage lead. We treated proposed valuations, IPO targets, and run-rate forecasts as targets, not results.
We treated Claude Code as evidence of professional monetization, not as a proxy for Claude’s total user base. Its role in the analysis comes from the revenue figures Anthropic disclosed for that product and from independent developer surveys.
We used third-party app-usage and web-traffic estimates only for directional consumer-scale comparison. Those estimates are useful for showing that ChatGPT remains much larger, but they are not equivalent to company-reported Claude-wide active users.
We separated revenue leadership from user leadership throughout the analysis. The main conclusion is not that Claude has more users than ChatGPT, but that Anthropic has monetized a smaller professional and enterprise base more densely.
Key sources used for this analysis include: OpenAI on its ARR series and compute constraints, OpenAI on ChatGPT weekly active users and consumer subscribers, OpenAI on the original ChatGPT launch, Anthropic on the Claude launch, Anthropic’s Series F announcement, Anthropic’s Series G announcement, Anthropic on its Google and Broadcom compute partnership, Anthropic on expanding Google Cloud TPU usage, Anthropic on Amazon compute expansion, AWS on Claude availability through Amazon Bedrock, Google Cloud on Claude availability through Vertex AI, Reuters, via The Star, on Anthropic’s IPO prospectus and partner dependence, Euronews on Anthropic’s third-quarter revenue and IPO plans, Axios on the gross-versus-net revenue gap, Bloomberg, via AI Weekly, on gross and net cloud-sales accounting, Similarweb on Claude and ChatGPT usage overlap and retention, JetBrains Research on AI coding-agent adoption, and Yahoo Finance quoting Sensor Tower app-usage estimates.
Written by the New Market Pitch teamWe track new markets so founders and investors can move faster. About us
