OpenAI says ChatGPT now reaches 1.2 billion people weekly, as its annualized revenue rate nears $70 billion, up about 70% since the start of Q3. Growth is driven by enterprise sales, the Codex coding assistant, and an aggressive price war, while Anthropic is close behind with a similar run rate.
SynthePulse Insight · AI deep reading
Version 1 · 1 source
At DevDay, OpenAI announced that ChatGPT's weekly active users exceed 1.2 billion, while its annualized revenue rate is reported to be approaching $70 billion. But both figures come from the company's own statements or media relay, and the real question is: can revenue growth cover the massive data center spending.
At DevDay, OpenAI announced that ChatGPT's weekly active users have exceeded 1.2 billion. At the same event, the company also disclosed that ChatGPT Work and Codex together have over 35 million weekly active users, and that the number of businesses using its products has reached 2.5 million. These figures all come from OpenAI's own announcements and lack independent third-party verification.
Alongside user scale, according to Axios, which first reported it, OpenAI's annualized revenue rate is approaching $70 billion, up about 70% from the start of the third quarter. To be clear, the annualized revenue rate (ARR) metric converts current monthly revenue into full-year revenue; it is a predictive metric, not revenue actually realized.
According to the same report, OpenAI's growth is driven by enterprise sales and an aggressive price war aimed at Claude and Chinese models. The company has just released GPT-6.1-Sol to double down on its price war strategy. The coding assistant Codex is also growing rapidly, riding the popularity of the GPT-6 model family.
These growth drivers are consistent in direction with the number of businesses announced at DevDay and the weekly active user data for ChatGPT Work and Codex, but the actual contribution of GPT-6.1-Sol and Codex growth to overall revenue is not clear. The long-term impact of the price war on enterprise sales and profit margins is likewise unclear.
According to reports, Anthropic's annualized revenue rate had already exceeded $65 billion in July, and it may now be on par with or ahead of OpenAI. Anthropic is preparing to launch an IPO as early as November.
Both pieces of information are "according to reports" and have not been confirmed by first parties, so whether Anthropic's annualized revenue rate has already caught up with or surpassed OpenAI is still uncertain, and whether its plan for an IPO as early as November can proceed on schedule is also uncertain.
Putting the above numbers together, the core question raised by the report is: can revenue grow fast enough to cover the massive data center bills. This is an inferential judgment, not a confirmed fact.
The practical constraints supporting this question are: all user and revenue figures come from OpenAI's DevDay announcements or Axios reporting and lack independent third-party verification; the ratio of paying users to free users among the 1.2 billion weekly active users is unclear; and the annualized revenue rate itself is a predictive metric. Under these constraints, whether revenue can cover data center spending is still unclear.
All user and revenue figures in this article come from OpenAI's DevDay announcements or Axios reporting and lack independent third-party verification. The information on Anthropic's revenue and IPO is "according to reports" and has not been confirmed by first parties. The annualized revenue rate (ARR) is a predictive metric that converts current monthly revenue into a full-year figure, not revenue actually realized. The core question "can revenue cover the data center bills" is an inferential judgment.
At DevDay, OpenAI gave figures including 1.2 billion weekly active users, an annualized revenue rate approaching $70 billion, and 2.5 million enterprise customers, with growth driven by enterprise sales and a price war aimed at Claude and Chinese models. But these figures lack independent verification, the annualized revenue rate itself is a predictive metric, and Anthropic's catch-up and IPO plans are only reported. The real unresolved question is not user scale, but whether revenue growth can catch up to the massive data center bills.
Secondary coverage
This article also covers this event; its overall headline is not the event headline.