OpenAI makes a dollar and loses two, while Anthropic has started turning a profit
Original author: Su Yang, Tencent Technology
In mid-May, two AI giants revealed their hands simultaneously—OpenAI secretly filed for an IPO, and Anthropic presented its first profitable quarter financial forecast.
Data shows that OpenAI generated $5.7 billion in revenue in the first quarter, but incurred a loss of $1.22 for every dollar earned. Anthropic, with $4.8 billion in revenue during the same period, lagged by nearly $1 billion, but its projected second-quarter growth rate surged, with revenue expected to reach $10.9 billion and an operating profit of approximately $559 million.
This difference gives the outside impression that one is a superstar valued at $1 trillion, still asking the market for patience, while the other is a former challenger that has quietly reached the threshold of profitability.
01. 5.7 billion vs 4.8 billion
A source familiar with the matter told The Information that OpenAI generated approximately $5.7 billion in revenue in the first quarter of this year, nearly $1 billion more than Anthropic’s $4.8 billion in revenue during the same period.
Looking at these two numbers alone, OpenAI's lead still appears significant.
OpenAI first-quarter financial results, source: The Information
According to the aforementioned insiders, the three main drivers behind OpenAI's first-quarter growth are the surge in popularity of the coding agent Codex, increased enterprise sales, and ChatGPT test advertising.
The surge of Codex highlights the strong demand among developers for tools that can get real work done, a need that overlaps significantly with Anthropic’s customer base. Meanwhile, the experimentation with advertising reveals OpenAI’s anxiety about finding monetization pathways within its large pool of free users.
OpenAI's average weekly active users in the first quarter were approximately 905 million, peaking at 920 million in February.
Growth begins to plateau when users reach an extremely large scale. Although it has increased its paid subscriber base to 55 million from 47 million at the end of last year, this represents a very low conversion rate compared to its over 900 million weekly active users.
Moreover, the inference cost associated with this portion represents a massive black hole for OpenAI.
On the other hand, Anthropic generated $4.8 billion in revenue in the first quarter, almost entirely from its core strength: selling AI models to enterprises and developers. It does not have a large free consumer base requiring massive subsidies, like ChatGPT. This difference may be a key factor in its potential to surpass its older competitor.
02. The Fastest Comeback in History
According to financial data disclosed by Anthropic to investors and obtained by The Wall Street Journal, the company expects revenue for the second quarter to reach $10.9 billion, more than doubling from the first quarter.
Moreover, its revenue growth rate has surpassed that of Google and Facebook before their IPOs.
Operating profit by business, source: Anthropic
The Information points out that by April 2026, Anthropic's annualized revenue exceeded $30 billion, while OpenAI's annualized revenue was approximately $25 billion.
At the 2026 May developer conference, Anthropic CEO Dario Amodei joked that their recent revenue growth had become so rapid it was “hard to handle.”
Anthropic's revenue growth, with red indicating the December 2025 forecast, source: The Information
Anthropic expects an operating profit of approximately $559 million in the second quarter, a milestone event. Last summer, the company shared projections with investors indicating it would not achieve full-year profitability until at least 2028.
However, operating profit excludes stock-based compensation expenses, and given the substantial future computing costs, Anthropic may not remain profitable throughout the entire fiscal year—but it demonstrates that AI model companies focused on enterprise customers can achieve a viable profit model in the short term.
In contrast, OpenAI, although its second-quarter expectations are not yet known, showed data presented to investors indicating that its first-quarter adjusted operating profit margin was -122%. In other words, for every dollar of revenue generated, the company lost $1.22.
OpenAI is not expected to achieve positive cash flow until 2029 or 2030, and before then, it needs to continuously fill a substantial funding gap.
HSBC analysts estimate that OpenAI faces a $207 billion funding gap relative to its growth plans. OpenAI CEO Sam Altman hinted at an all-hands meeting that even after filing IPO documents, the actual listing could be delayed, as filing an application is "not the same as being ready to go public."
The financial pressure behind this is obvious.
03. The Same AI, Two Different Fates
Why have the financial situations of two companies diverged so sharply amid the same AI wave?
The answer lies within different client structures.
According to Forbes' analysis, approximately 85% of Anthropic's revenue comes from enterprise and developer customers. More than 500 companies spend over $1 million annually on the Claude platform, and 8 of the Fortune 10 companies are its clients.
Enterprise customers have clear willingness to pay, more predictable usage patterns, lower service costs, and more sticky contracts. This is a healthy, sustainable business model.
In the first quarter, Anthropic spent 71 cents on compute for every dollar of revenue earned, and this is expected to drop to 56 cents in the second quarter, demonstrating an immediate improvement in efficiency.
In contrast, OpenAI generates approximately 85% of its revenue from ChatGPT consumer subscriptions. Despite having 55 million paid subscribers, it supports over 900 million weekly active users who do not contribute corresponding revenue, resulting in a structural deficit.
OpenAI is not unaware of this.
Under the leadership of executives such as CEO Fidji Simo, the company has begun cutting costly projects like the video generation app Sora, aiming to refocus on revenue-generating businesses and commercial clients. However, turning around a business model centered on free consumers is no small feat.
Of course, directly comparing the revenue figures of the two companies requires accounting for a key difference in accounting treatment.
The Information explains this in detail: Anthropic records all technical sales made through cloud partners such as Amazon and Google as full revenue. In contrast, OpenAI, due to its long-term special partnership with Microsoft, in which Microsoft holds exclusive rights to use its intellectual property, recognizes only 20% of the revenue from model sales via Microsoft Azure as its own income.
However, note that the two companies have slightly different accounting practices, and both report somewhat inflated revenues: Anthropic includes the full revenue from cloud providers like Amazon and Google reselling its models, without deducting revenue shares; OpenAI, which must distribute 20% of its revenue to Microsoft before 2030 (potentially reaching $6 billion this year), does not report any sales generated through cloud partners at all.
However, even if OpenAI adopted Anthropic’s approach to increase its annual revenue by billions of dollars, it would still fall short of bridging the hundreds-of-billions-of-dollars gap between them.
04. Behind the IPO Race
On the path to an IPO, all financial secrets will be brought into the light.
OpenAI, Anthropic, and Elon Musk’s SpaceX are all racing to go public, with each company’s valuation potentially exceeding $1 trillion.
Currently, OpenAI has secured $122 billion in funding from suppliers such as Amazon and NVIDIA and is seeking to go public as early as September 2026. Meanwhile, Anthropic is continuing a funding round that could value it above OpenAI and is considering an IPO as early as October. Altman has privately expressed his desire to go public first.
Anthropic now holds a quarter of data that has proven to be profitable.
Even if future losses recur due to astronomical investments in computing infrastructure—such as paying $1.25 billion monthly to SpaceX for data center capacity and new large-scale compute contracts with Broadcom and Google—it has still proven to the market that its business model works. Its story is that of an enterprise software company, comparable to Salesforce or ServiceNow.
OpenAI, on the other hand, presented public market investors with a story requiring greater conviction—it needed to convince the market that AI agents, image generation, and a vast future advertising business would ultimately convert its massive consumer traffic into profits.
According to Otman's projections, ChatGPT's advertising business could generate approximately $102 billion in revenue by 2030.
But this will take time, and time is precisely what OpenAI lacks as it trades losses for growth.
OpenAI has just launched over two gigawatts of computing power, surpassing the total of SpaceX’s entire Colossus cluster, and all of this requires funding.
So for investors, when the S-1 filing is made public, should we believe in a company that has already found a profitable model, or a giant asking the market for more years and hundreds of billions of dollars to explore potential profitability? The answer will determine the fate of both companies.
Original link
Click to learn about the open positions at BlockBeats
Welcome to the official BlockBeats community:
Telegram subscription group: https://t.me/theblockbeats
Telegram group: https://t.me/BlockBeats_App
Official Twitter account: https://twitter.com/BlockBeatsAsia