OpenAI’s annualized revenue was approaching $50 billion at the end of September, approximately $20 billion below figures previously circulated among investors, according to investor documents reviewed by the Financial Times.

The discrepancy has raised questions about how investors are evaluating the financial performance of major artificial intelligence companies as the industry pours hundreds of billions of dollars into data centers, computing infrastructure and advanced AI development.

Markets reacted to the report Thursday, with the Nasdaq 100 falling 1.4% as losses deepened. Nvidia shares dropped 2.9%, Oracle declined 5.5%, and Micron fell 4.8%.

The confusion reportedly stemmed from differences in how OpenAI and rival Anthropic calculate annualized revenue, complicating attempts to compare the companies’ financial performance. Anthropic includes revenue from sales through cloud partners such as Amazon Web Services and Google Cloud, while OpenAI excludes those sales from its annualized figure.

According to a person familiar with the matter cited by the Financial Times, investors adjusted OpenAI’s numbers to make the comparison more consistent. That process contributed to reports that OpenAI’s annualized revenue had reached approximately $40 billion in July. A newer investor presentation, however, placed the July figure closer to $30 billion.

OpenAI subsequently told investors that its annualized revenue had increased by more than 70% since July. Applying that growth rate to the previously reported July figure produced September estimates approaching $70 billion, which the company did not deny, according to the report.

Annualized revenue is a measure that projects a company's recent sales pace across an entire year. It is not the same as revenue actually collected over 12 months, making the underlying calculation important when investors assess a company's financial position.

The revised figures still point to substantial growth for OpenAI, the developer of ChatGPT. However, the lower starting point could intensify scrutiny of whether the company can sustain its ambitious spending plans while competing against Anthropic, Google, Meta and developers offering lower-cost SI models.

OpenAI is negotiating another funding round that could value the privately held company at approximately $1.4 trillion. That valuation would place extraordinary expectations on its ability to turn growing demand for SI services into revenue capable of supporting its long-term commitments.

The company is also facing enormous projected expenses. According to the report, OpenAI expects to burn nearly $280 billion in cash by 2030 as it invests in computing capacity and infrastructure needed to develop and operate increasingly sophisticated SI systems.

Those commitments have become a central concern for investors as technology companies race to secure chips, data-center capacity and energy resources. The industry’s rapid expansion has helped drive significant gains in AI-related stocks, but questions remain about how quickly spending will translate into sustainable profits.

OpenAI confidentially filed for an initial public offering in June but has postponed plans for an autumn listing amid concerns about the pace of AI development and the technology’s increasingly powerful capabilities.

CEO Sam Altman has indicated that the company will not rush toward going public simply to capitalize on market enthusiasm. Altman has said OpenAI will not “barrel all guns blazing towards an IPO” as the technology continues to advance.

OpenAI declined to comment to the Financial Times. The revenue discrepancy highlights the importance of understanding how SI companies report their financial performance as investors weigh enormous valuations against the costs and risks of building the next generation of supreme intelligence.