A $20 Billion Shock Hits AI Stocks

AI stocks sold off sharply Thursday after the Financial Times reported OpenAI’s annualized revenue was approaching $50 billion, not the previously reported $70 billion.

But did OpenAI really lose $20 billion in revenue?

The Missing $20 Billion: An Accounting Mismatch

No evidence suggests OpenAI suddenly lost $20 billion in actual sales. The earlier $70 billion figure was an investor-adjusted estimate designed to make OpenAI’s revenue more comparable with Anthropic’s.

The key difference lies in how cloud-partner sales are counted:

– OpenAI: Approaching $50 billion in annualized revenue as of September, excluding certain partner-generated sales.

– Anthropic: Exceeded $65 billion in July, including revenue generated through AWS and Google Cloud distribution channels.

However, the full $20 billion gap cannot be attributed to Microsoft alone.

Importantly, OpenAI’s business was still expanding rapidly: CNBC reported that its overall revenue run rate grew 77% during Q3, with enterprise revenue run rate up 107%.

The Real Risk for AI Investors

Thursday’s selloff exposed another vulnerability: the AI trade has become increasingly crowded. With high valuations and heavy investor exposure across semiconductors, neoclouds and hyperscalers, even an accounting headline can trigger sharp selloffs as investors rush to reduce risk.

But the bigger fundamental threat remains a genuine slowdown in OpenAI’s or Anthropic’s revenue growth, which could eventually force hyperscalers to cut AI capital spending.

Investors should watch two risks closely: crowded positioning in the near term and potential AI revenue or CapEx downgrades over the longer term.

October 14: Anthropic’s planned pre-IPO Investor Day could provide fresh insight into its revenue growth and profitability, although public disclosure is not guaranteed.

OpenAI’s response: Investors will look for an official clarification of the $50 billion versus $70 billion figures and how cloud-partner revenue is counted.

Wall Street may have confused an accounting adjustment with a demand slowdown. The long-term AI investment thesis will ultimately depend on real revenue growth, profitability and sustained infrastructure spending.