Retail Investors Piled Into Chip Stocks. Now OpenAI’s $50 Billion Revenue Shock Is Spreading Fear

OpenAI Revenue Disappoints, Dousing AI Optimism

Annualized Revenue 29% Below Expectations at $50 Billion

OpenAI’s annualized revenue is about $50 billion, roughly $20 billion below market expectations. The shortfall has fueled fresh doubts about the growth outlook for the artificial intelligence industry and sent AI infrastructure stocks lower.

The Financial Times reported on Oct. 8 that OpenAI told investors its annualized revenue was about $50 billion. That is 29% below the $70 billion the market had been expecting. Annualized revenue assumes sales from a given period continue for a full year.

The weaker-than-expected figure undercut bullish sentiment on AI. Shares of AI chip companies including Nvidia fell that day.

‘Cracks in the AI Growth Story’ vs. ‘Just a Difference in Accounting’

$50 Billion, Not the Estimated $70 Billion, Sends Tech Stocks Lower on AI Growth Concerns

Confusion over OpenAI’s annualized revenue dates back two or three months. As interest grew in the company’s results ahead of a planned initial public offering, remarks by its chief financial officer at a private event about rising revenue filtered into the market. Estimates were then layered on top, producing the $70 billion figure that emerged last month. OpenAI’s officially tallied annualized revenue, however, was about $50 billion as of late last month. The gap has raised concern that optimism about the AI industry could weaken.

Opaque Revenue Calculation Structure

On Oct. 8, the Financial Times said the confusion began as investors compared OpenAI’s performance with rival Anthropic. Annualized revenue refers to yearly revenue calculated by assuming results from a specific period continue for the next 12 months. If a company generated $1 billion in revenue in September, for example, its annualized revenue would be $12 billion.

Annualized revenue is a key gauge of growth in the AI industry. But the two companies calculate it differently. When Anthropic provides AI services through cloud providers such as Amazon Web Services and Microsoft, it includes the full amount users pay those platforms as revenue. Fees paid to the cloud companies are treated as costs. OpenAI counts only the portion of partner sales that it actually receives as revenue.

That difference appears to have driven the confusion. The Financial Times said investors trying to compare OpenAI’s revenue using Anthropic’s method circulated a claim that OpenAI’s annualized revenue was $40 billion as of July, when the actual figure was $30 billion. A separate comment attributed to an OpenAI official that revenue had risen 70% from July then helped fuel the $70 billion estimate.

Ripple Effects May Reach the AI Supply Chain

The market reaction was sharp. Investors had been operating on the assumption that OpenAI’s annualized revenue was $70 billion. The new figure effectively cut that by $20 billion, or 29%.

Debate could also intensify over Anthropic’s approach to revenue recognition. The issue is whether it is reasonable to book the full value of sales made through cloud partners as revenue. According to the Financial Times, Anthropic’s annualized revenue stood at $6.5 billion as of July. If it were calculated the way OpenAI reports revenue, net of fees, Anthropic’s annualized revenue would probably be lower.

The news also hit AI bullishness more broadly. Nvidia, the leading AI chip stock, fell 2.94%. Broadcom dropped 4.35%, Micron Technology slid 4.79%, Advanced Micro Devices lost 3.90%, and SK Hynix ADRs declined 4.35%, dragging the Nasdaq lower. If actual results at OpenAI or Anthropic fall short of expectations, spending on semiconductors and data centers may also come in below forecasts. That could weigh on the lofty valuations and growth expectations attached to AI companies.

Experts are divided. Gil Luria, an analyst at D.A. Davidson, said the confusion stemmed less from OpenAI’s revenue actually falling from $70 billion to $50 billion than from comparisons between different revenue-recognition methods. More important than the headline figure, he added, is the pace of the underlying business.

Others urged caution. Ross Mayfield, an investment strategist at Baird, said that if the news creates cracks in the AI growth narrative by showing OpenAI generated less revenue than expected, the impact could spread across the AI supply chain. This development will clearly have a broad effect on the industry, he said.

Hwang Jeong-su, New York correspondent, Korea Economic Daily, hjs@hankyung.com