OpenAI is making money at a rate that most businesses could only imagine. Wall Street continued to have cause for concern.
The ChatGPT maker recently told investors that its annualized revenue was approaching $50 billion at the end of September, roughly $20 billion below the nearly $70 billion figure that had been widely reported, the Financial Times reported.
The discrepancy doesn’t seem to indicate that OpenAI lost $20 billion in revenue overnight. Reuters claims that it primarily reflects differences in how rival Anthropic and OpenAI calculate revenue. Whereas OpenAI excludes some revenue from cloud partners, Anthropic doesn’t.
However, investors responded swiftly.
Since the report raised questions about whether AI applications’ revenue can justify the huge sums being spent on chips, data centers, and financing, Nvidia (NVDA), Oracle (ORCL), and other AI infrastructure boom companies fell on Oct. 8.
That is what makes the discrepancy important.
One of the IT startups that is expanding the quickest is OpenAI. However, it is also in the middle of a cycle of investments worth hundreds of billions of dollars.
Wall Street takes notice when the revenue figure sustaining that ecosystem abruptly declines.
OpenAI’s $20 billion gap is largely about how revenue is counted
Investors should first realize that OpenAI did not announce a traditional $50 billion yearly sales number.
An annualized revenue run rate was provided.
To calculate that statistic, income from a recent month is usually multiplied by 12. It may be helpful for gauging a private firm that is expanding quickly, but when growth is erratic or businesses compute the measure differently, it can also give an inaccurate impression.
This seems to be the case.
At the end of September, OpenAI told investors that its annual revenue was almost $50 billion, according to Reuters, which verified the Financial Times article. The approximately $70 billion figure that was previously circulating among investors reflected an attempt to make OpenAI’s figures similar to Anthropic’s.
In its run-rate estimates, Anthropic takes into account income from cloud partners such as Alphabet’s (GOOGL) Google Cloud and Amazon’s (AMZN) Amazon Web Services. OpenAI does not include comparable partner sales in the same way.
When billions of dollars pass via cloud platforms, that disparity may grow significantly.
About half of Anthropic’s revenue last year came through cloud partners, who took about 16% of every dollar generated through those relationships, Reuters reported.
Additionally, it appears that investor efforts to harmonize those disparate techniques led to the earlier OpenAI projection. According to The Information, OpenAI CFO Sarah Friar’s remarks that the company’s revenue run rate had climbed by around 70% between the end of June and the end of September prompted the higher forecasts.
That means the $20 billion gap is more complicated than an ordinary revenue miss.
But markets rarely wait for accounting debates to be settled before reacting.
AI stocks show why one OpenAI number matters
Although OpenAI is a private company, investors may still indirectly trade its performance via the businesses constructing the infrastructure it requires.
The revenue data swiftly leaked into public markets as a result.
More Artificial Intelligence:
- AI’s biggest business opportunity may not look like AI
- AI racks are outgrowing old power: 2 chipmakers have a strategy
- BofA splits software into winners and losers as AI agents arrive
Nvidia fell 2.9% on Oct. 8, while Oracle dropped 5.5% and Micron Technology (MU) declined 4.8%, according to MarketWatch. Advanced Micro Devices (AMD), Broadcom (AVGO), and Intel (INTC) also fell sharply.
Semiconductor companies were among the worst segments of the market, and the Nasdaq Composite closed the afternoon down almost 1.3%.
When compared to the magnitude of the promises being made around OpenAI, the response makes more sense.
As technology companies race to buy chips and expand AI infrastructure, Broadcom and Oracle are seeking $50 billion in OpenAI financing, Reuters reported.
For investors, it raises a straightforward question.
That infrastructure may continue to be widely used and provide profitable returns if OpenAI and other AI developers continue to expand quickly.
Businesses across the supply chain may find themselves battling for less money than their investment plans anticipated if revenue growth finally falls short of expenditure.
The $50 billion number does not establish that second possibility. OpenAI continues to expand at a very rapid pace.
However, the market’s response demonstrated how sensitive investors have become to anything that casts doubt on their presumptions about the amount of money at the end of the AI expenditure chain.
OpenAI is still growing at an extraordinary pace
There is another side to the story that could easily get lost in the $20 billion headline: The real growth of OpenAI is still huge.
The company entered 2026 with roughly $20 billion in annualized revenue, compared with about $6 billion in 2024. By the end of September, that figure was approaching $50 billion.
Even with the lower number, the figure indicates that the company’s run rate has more than doubled since the start of the year.
In other words, the recent revelation does not prove that consumers have suddenly stopped purchasing ChatGPT or OpenAI’s business offerings.
It is proof that investors should use caution when comparing private AI firms using financial indicators that may not be determined using the same methodology.
With OpenAI and Anthropic getting closer to public markets, that problem is becoming more crucial.
Both companies are preparing for eventual initial public offerings, according to Reuters. Instead of mainly depending on privately distributed run-rate data, going public would provide investors with standardized financial statements and subject their finances to more scrutiny.
Additionally, the two businesses are becoming more competitive.
OpenAI generated about $6.7 billion of quarterly revenue in the second quarter, while Anthropic generated approximately $11.5 billion, Reuters reported. Anthropic’s annualized revenue passed $65 billion in July.
These figures highlight how rapidly the market is evolving. Only a few years ago, these businesses were relatively modest startups. Today, they generate income at levels comparable to giant public organizations, all the while consuming enormous sums of cash.
Financial openness becomes more vital as a result of that combination.
Investors are keen to know more than whether AI is useful or whether consumers want to use it. They are becoming more interested in learning how much it costs to generate that demand and who ultimately benefits.
OpenAI’s revenue gap raises a bigger question about AI spending
In the end, the market could conclude that it overreacted to the OpenAI study.
Rather than revealing an abrupt decline in OpenAI’s operations, the discrepancy between approximately $50 billion and $70 billion seems to stem mostly from the presentation and comparison of revenue.
However, the sell-off exposed a crucial aspect of the AI trade.
For several years, indications of strong demand were often sufficient to propel AI-related stocks upward. The question of whether that demand can generate enough revenue to fund the infrastructure being developed around it is now becoming more important to investors.
Customers must continue ordering processors from chipmakers. For cloud enterprises to be active, those chips are essential. Tenants that are prepared to sign lengthy contracts are essential to data center owners. Businesses that take out loans to fund AI infrastructure must generate enough income to pay back the debt.
Companies such as OpenAI and Anthropic are at the end of a large portion of that chain.
The roughly $50 billion run rate of OpenAI indicates that demand is still forceful. However, the response to a $20 billion disparity demonstrates how much faith is already ingrained in the larger AI investment narrative.
That could be the disclosure’s most important lesson. The average investor is increasingly concerned about whether income will rise fast enough to support everything Wall Street is presently funding.
Related: OpenAI paused training again, and Washington’s in the middle of it