The OpenAI Revenue Revision: Unpacking the $50B vs. $70B Debate

A recent report by the Financial Times has clarified a significant discrepancy in the perceived financial performance of OpenAI. According to the paper, OpenAI disclosed to investors that its annualized revenue stood near $50 billion as of the end of September. This figure starkly contrasts with the $70 billion annualized run rate widely reported by the FT and other media outlets just last month.

How Did the $70 Billion Figure Emerge?

The previously circulated $70 billion estimate was the product of investor calculations based on fragmented data. Sources indicate that investors first adjusted OpenAI's revenue accounting to arrive at an August annualized run rate of $40 billion. They then extrapolated from the company's disclosed information—specifically that revenue had grown more than 70%—to project the $70 billion figure.

This process highlights a common market practice in the absence of full transparency: constructing growth narratives from incomplete data points. However, such extrapolation overlooked a critical factor—whether the base figures and accounting methodologies were consistent.

The Core Issue: A Case of Comparing Apples to Oranges

The root of the confusion, according to people familiar with the matter, lies in investors' attempts to directly compare OpenAI's metrics with those of its key rival, Anthropic. The fundamental problem is that the two companies employ different revenue recognition principles.

  • Anthropic's Approach: Includes the full sales revenue generated through cloud service partners like Amazon AWS and Google Cloud in its top-line figures. This "gross flow" accounting can rapidly inflate reported revenue.
  • OpenAI's Approach: Is more conservative. It does not incorporate revenue from cloud partnerships into its own reported revenue. This suggests its disclosed numbers are closer to "net" or core business revenue.

Comparing these two methodologies is akin to comparing a brand's total retail sales with its wholesale factory revenue—a direct comparison yields misleading conclusions. When investors applied Anthropic's broader accounting standard to OpenAI, they inadvertently created a perception of a massive growth gap and expectation mismatch.

Market Implications and Industry Reflection

This incident goes beyond correcting a single number. It reflects the transparency challenges within the AI industry amid frenzied fundraising and valuation contests. The limited financial disclosure from private companies leaves ample room for market interpretation and speculation, often fueling narratives based on partial information.

For OpenAI, an annualized run rate approaching $50 billion remains a significant milestone, underscoring its strong commercial momentum. Clarifying this discrepancy may allow the market to better understand the true quality and trajectory of its revenue. OpenAI has declined to comment on the report.

The "accounting standards debate" serves as a reminder to investors and observers: when assessing the financial health of these tech giants, understanding the logic behind the numbers is as crucial as the figures themselves. In the next phase of the AI race, solid revenue quality may hold more long-term significance than impressive top-line volume.