OpenAI Misled Everyone

It seems that too many supposedly-sensible people believe that the information given out by large companies can be taken at face value. We wonder, dear readers, how it came to be that the populace has slowly lost its sense of skepticism for such matters; to any individual who has learned their history well, such a stance should seem quite odd.

Consider, for example, the situation of MicroStrategy in the late nineties. This company is credited, somewhat infamously, with popping the dot com bubble in 2000 due to a “restatement” of its financials for the previous two years which turned a profit into a loss. This was largely due to a result of it changing how it reported its revenue, which was inflated by circular deals in which the company sold its software to the same company from which it bought services.

We know well that there are lies, damn lies, and statistics, of course, and this is why the American financial world adheres to so-called GAAP, that is, Generally Accepted Accounting Principles. It may behoove you to know that these principles evolved out of the Great Depression, where financial chicanery all but broke the US economy for more than a decade; the lessons taught therein have been remarkably persistent in spite of late stage capitalism’s constant chipping away at them. To adhere to GAAP is a hard requirement for public companies when reporting their financial position, even though they may report non-GAAP metrics as well. This is no protection from bad data, however, as MicroStrategy was indeed a public company during its dot com era in spite of its financial shenanigans.

Now, having said all this, let us turn to OpenAI.

As it is currently a private company, it has no obligation to follow GAAP, and such we have some… rather nebulous data points with which to guide our assessment of its financial situation. Its favorite metric is annualized revenue or annual recurring revenue, and it is quite fond of touting these numbers, jazz hands-style, in front of every media outlet from here to London. Said media outlets dutifully report these figures without an ounce of suspicion or even explanation, most of the time. (Such breathless sycophancy is contemptible, in our most enlightened opinion.)

Annualized revenue is a somewhat interesting metric, used for operational analysis, but should not in any way be a substitute for actual, real revenue. It is most useful when examining repeated and reliable revenue, such as from subscriptions, and you may assume that it works for OpenAI, with its subscription-based business model. Unfortunately, you would be wrong, as OpenAI’s revenue is anything but reliable. The proof is in the pudding, as they say, and OpenAI’s pudding as been proven to be lacking in filling.

CFO Sarah Friar stated that OpenAI’s annualized revenue for 2025 was USD$20 billion, up from USD$6 billion in 2024. An exciting jump, at first glance, and it was certainly appreciated when it was reported in January of this year. However, the esteemed Ed Zitron reported the real financial metrics of OpenAI for those same years, metrics which were independently verified by the Financial Times and told a far more twisted tale of a company bleeding billions of dollars with no end in sight.

Its actual revenue for 2024 was USD$3.7 billion, not USD$6 billion.

Its actual revenue for 2025 was USD$13.07 billion, not USD$20 billion.

How, then, can we trust anything reported as annualized revenue from OpenAI? The disparity is enormous! It is, in our vaunted opinion, a financial sleight-of-hand designed for the sole purpose of misleading the public into thinking OpenAI is doing better than it truly is. As of March of this year, the adoring media reported that its 2026 annualized revenue would reach USD$25 billion; we may surely assume that this figure is utter nonsense, and (based on the trend from previous years alone) the actual 2026 revenue will be circa USD$16 billion, if anything.

This is a damning and horrifying indictment of the current media landscape. Every outlet which reported on OpenAI’s annualized revenue without question should be ashamed of their delinquency; they have entirely shirked their duty to question, discover, and inform. Any further reporting of such metrics, without mention of the real figures, should be treated with the utmost disdain.