OpenAI is facing scrutiny over a $20 billion revenue calculation controversy. On September 29, Axios, citing investor sources, reported that OpenAI's annualized revenue was approaching $70 billion. However, on October 8, the Financial Times revealed that OpenAI's annualized revenue presented to investors at the end of September was actually only close to $50 billion, a difference of approximately $20 billion, equivalent to inflating the revenue figure by 40%.
Axios subsequently explained that the $70 billion figure was not OpenAI's original disclosure, but rather a result recalculated by investors using a similar gross method to compare with competitor Anthropic. Anthropic includes sales from cloud partners such as AWS and Google Cloud in its revenue, then lists revenue sharing as an expense; OpenAI, on the other hand, uses a net method for some partner sales, recognizing only its own revenue.
Ironically, OpenAI had previously criticized Anthropic's accounting methods for exaggerating revenue, but now investors are using similar methods to inflate OpenAI's growth performance. Meanwhile, Bloomberg reported that OpenAI still expects annualized revenue of $70 billion by the end of the year and is seeking at least $30 billion in funding, with a pre-money valuation in discussions reaching as high as $1.4 trillion. It's important to emphasize that annualized revenue is not actual annual revenue, much less profit.
Currently, there is no evidence that OpenAI falsified its financial statements, but this controversy highlights a problem in the AI funding market: the same business can present drastically different growth figures depending on the statistical methods used, influencing investors' judgments about valuation and funding needs.
