OpenAI’s revenue math is giving Wall Street a headache

OpenAI’s annualized revenue run rate stood at approximately $50 billion at the end of September, according to the Financial Times. That would be an eye-popping number for almost any company on Earth.

The trouble is that investors had been passing around a figure closer to $70 billion in late September. When your audience expects a seven and you hand them a five, the reaction is predictable.

The gap did not come from customers walking away. It came from two different ways of counting the same money. That distinction matters a lot when a company is reportedly eyeing a public listing at a valuation of around $1.4 trillion.

Gross versus net: the accounting fork in the road

Picture a travel agent who sells a $1,000 flight and keeps a $100 commission. One way of booking that sale records $1,000 in revenue. The other records only the $100 the agent actually keeps.

That, roughly, is the split at the heart of this story. Per the research findings, OpenAI reports revenue on a net basis. Investors, however, had been applying a gross methodology similar to the one used by rival Anthropic, which includes sales made through partners.

The FT report landed on October 8, 2026, and it quickly reshaped how the market was reading OpenAI’s momentum.

OpenAI has pushed back on any suggestion that business is slowing. The company announced 77% overall run-rate growth in the third quarter of 2026. Enterprise revenue did even better, growing 107% over the same stretch.

The company remains optimistic about reaching or topping roughly $70 billion in annualized revenue by year end, this time on its own net accounting terms.

A year of fast climbs and a few stumbles

The trajectory in 2026 has been steep. OpenAI started the year with an annualized revenue run rate of approximately $20 billion. By August, it had passed $40 billion.

Reaching approximately $50 billion by the end of September keeps that curve pointing up. Hitting at least $70 billion by December would mean more than tripling the run rate in a single calendar year.

The year has not been a straight line, though. In April 2026, OpenAI reportedly fell short of internal targets for ChatGPT user growth and some revenue benchmarks. That stretch came as competition intensified from Google and Anthropic.

Meanwhile, OpenAI’s spending ambitions remain enormous. The company has leaned into massive capital expenditures and infrastructure deals with tech giants including Microsoft and Amazon.

What this means for investors and the AI trade

The most immediate fallout showed up outside OpenAI. Oracle shares fell 5% on October 8 as the run-rate news circulated.

For investors, the lesson is less about whether OpenAI is growing and more about how growth gets reported. Comparing companies that use different revenue methods is a bit like comparing a marathon time to a half-marathon time. Both numbers are real, but the comparison is meaningless without context.

There is also a competitive wrinkle. If Anthropic’s gross approach produces larger headline numbers, casual comparisons may flatter one company over another in terms of who is winning the enterprise race.

The enterprise figures may be the most telling detail in OpenAI’s update. Growth of 107% in that segment suggests businesses are paying up for its tools, which tends to be stickier revenue than consumer subscriptions.

What to watch next is whether OpenAI actually reaches its year-end target of at least $70 billion on a net basis. Clearing that bar would go a long way toward quieting the accounting debate. Falling short would likely revive it, and given the Oracle reaction, the ripples would not stay contained to one company.

Investors should also keep an eye on how AI companies describe their revenue going forward. Until then, anyone sizing up an AI company’s growth would be wise to ask a simple question first: gross or net?