The missing $20 billion was not lost revenue, but the accounting gap was enough to unsettle markets.
San Francisco, United States
OpenAI has told investors that its annualized revenue was approaching $50 billion at the end of September, significantly below the roughly $70 billion figure that had circulated in recent reports. The difference immediately reverberated through technology markets, intensifying scrutiny over how the fastest-growing artificial intelligence companies measure revenue and justify extraordinary levels of investment.
The gap does not mean OpenAI suddenly lost $20 billion in sales. The higher figure was based on estimates that adjusted OpenAI’s business to make it more comparable with rival Anthropic, which includes certain customer spending through cloud platforms in its revenue calculations. OpenAI excludes some of those partner-channel sales, creating a substantial difference in how the two companies are measured.
That distinction matters because annualized revenue is not the same as audited yearly revenue. It extrapolates recent sales over a full year, making it useful for fast-growing companies but also vulnerable to distortion when growth rates, accounting methods or distribution structures differ. For AI companies expanding at unprecedented speed, seemingly small methodological choices can translate into tens of billions of dollars.
Markets reacted sharply. Technology stocks fell as investors reassessed whether the explosive growth expected from generative AI can keep pace with the enormous infrastructure commitments being made across the industry. Companies heavily exposed to OpenAI’s expansion, including cloud providers, chipmakers and data-center operators, came under particular pressure.
The broader concern is not simply whether OpenAI is growing. Its revenue expansion remains exceptionally rapid. The deeper question is whether revenue can eventually justify the capital required to train increasingly powerful models, purchase advanced chips, build data centers and secure enough electricity to operate them.
That challenge extends far beyond one company. The AI industry is entering a phase in which enthusiasm about technological capability is being tested against cash flow, financing costs and infrastructure economics. Investors increasingly want evidence that unprecedented spending on compute will translate into durable profits rather than permanent capital requirements.
At the same time, demand for AI hardware remains strong. Semiconductor manufacturers and foundries continue reporting record or near-record activity, suggesting that companies are still accelerating investment in artificial intelligence despite growing financial skepticism.
OpenAI is also reportedly preparing another major fundraising effort that could push its valuation substantially higher. Any future public listing would bring considerably greater transparency because audited financial statements would allow investors to examine revenue, losses, infrastructure commitments and cash consumption more directly.
The current episode therefore represents something larger than a disputed revenue number. Artificial intelligence is moving from an era dominated by technological possibility into one increasingly governed by financial accountability.
The AI race is no longer asking only how powerful the models can become. Markets are beginning to ask whether the economics can become powerful enough to sustain them.