OpenAI Is In A Far Worse Position Than Anyone Thought It Was

This is damning.

Photo by Jp Valery on Unsplash

Everyone with even a modicum of sense has been predicting OpenAI’s self-inflicted demise for a while now. It is painfully obvious that this is not a sustainable business. Their product is insanely expensive and has basically no economies of scale, and while it might be impressive to a specific breed of narcissists and sycophants, it simply doesn’t bring in enough cash to cover its ludicrous operational costs. OpenAI is more akin to a cash-hungry black hole than a sustainable business. Many, including myself, have predicted that it will go bankrupt soon (read more here). But it turns out OpenAI’s ship is sinking significantly faster than even the most pessimistic predicted. Let me explain.

This revelation comes from Ed Zitron, who received an exclusive look at OpenAI’s IPO filings, which detailed the company’s actual balance sheet for 2024 and 2025. But before we dive into this information, to truly wrap our heads around the news, we need some context. Currently, OpenAI is a private company, which means that we don’t have a clear view of its financials. So, analysts have had to estimate its balance sheets based on Sam Altman’s breadcrumb clues. Would you like to hear people’s predictions for OpenAI’s 2025 financials?

On the very optimistic end, one external report predicted that OpenAI would post a net loss of $8 billion in 2025. That is a major escalation from 2024, in which they posted a $5 billion loss and had to effectively be bailed out by their corporate backers. I cannot emphasise this point enough, but an $8 billion loss would still be horrific news for OpenAI, as it proves their losses grow as they scale, not shrink. That heavily implies that, no matter how much they grow, they will always burn cash.

On the more realistic end, others predicted 2025 would be significantly worse for Altman. Take my previous article, in which I used analysis from The Information, which predicted that OpenAI would have operating costs of $28 billion in 2025 and a grounded estimate of $11.9 billion in revenue for the year, resulting in an estimated net loss of $15.6 billion. That is nearly double the previous prediction. This is actually higher than some predictions for their 2026 net loss ($14 billion), and these predictions also maintained OpenAI would go bankrupt by 2027. That should give you an idea of the horrific consequences involved in such a colossal loss.

However, other estimates painted a much more grim outlook. I covered The Information’s reporting on OpenAI’s 2025 H1 financials in my previous article. In the first half of 2025, OpenAI reportedly generated $4.3 billion in revenue while incurring $17.8 billion in operating costs, resulting in a net loss of $13.5 billion. That meant OpenAI was losing about three times more money than it was earning and was on track to post a net loss of roughly $27 billion by the end of the year.

So, in the first half of 2025, OpenAI had already exceeded the more optimistic $8 billion annual loss prediction and nearly met the $15.6 billion annual loss prediction in just six months! Their six-month loss was also triple their posted loss for the entire year of 2024, demonstrating a major escalation.

Now we understand the context, we can talk about Ed Zitron’s exclusive reporting.

Zitron was able to get a hold of audited financial documents for OpenAI’s upcoming IPO that have been independently verified by the FT, which detail the company’s actual finances for 2024 and 2025.

These documents verified that OpenAI did indeed incur $5 billion in losses in 2024. But they also revealed that OpenAI lost $38.5 billion of $13.07 billion in revenue in 2025. That is a nearly eightfold increase from 2024 and blows past even my most pessimistic predictions by more than $10 billion.

Admittedly, there is some nuance to that $38.5 billion figure, given that audited accounting is a lot messier than you might think. As more information comes out, I and many others will delve further into that nuance. However, $38.5 billion is the more realistic overall figure.

As Zitron put it, “$38.53 billion in losses are astronomical and far higher than most believed it would be. Losses also appear to be mounting year-over-year at a dramatic rate, and I’m not sure how this company finds a way toward any kind of sustainability or profitability.”

I couldn’t agree more.

There is zero evidence that any AI company is inference-positive, where the cost of querying the AI exceeds the amount the customer pays. As such, the larger these companies scale, the deeper their losses become, even if they completely disregard the crippling costs of training these AIs. OpenAI’s skyrocketing losses support this theory and strongly suggest that scaling is not a viable path to profitability.

But there is also the fact that AI has to be sold at a loss to attract customers. Take the recent report that discovered AI is more expensive than the workers it is intended to replace, even when it is being sold at an astronomical loss (read more here). So, even if OpenAI charged enough to cover their costs, their customer base would collapse, causing their revenue to collapse, and their losses would skyrocket.

Okay, so what does this mean for the future?

Well, Ed Zitron has said that he will release a much more in-depth analysis of these documents soon, so if you’re interested in reading that, go and subscribe to him. I am sure some part of it will cover what this means going forward.

But we already have some hints as to just how bad 2026 is shaping up to be for OpenAI.

The Information recently released a report on OpenAI’s 2026 Q1 financials. In this quarter, OpenAI incurred a $6.95 billion non-GAAP loss on $5.7 billion in revenue. GAAP stands for ‘Generally Accepted Accounting Principles’ and is the standard for clear and honest accounting. As such, non-GAAP financials can be very deceptive, given that expenses such as restructuring costs and certain forms of depreciation can be completely excluded. Because of this, their actual loss for the quarter is almost certainly much higher than $6.95 billion.

This isn’t great, considering that if nothing changes over the remaining three quarters, OpenAI would post a non-GAAP loss of $27.8 billion for 2026. But we know that OpenAI is losing customers to Anthropic and debating aggressively slashing prices to win them back, while its projected $50 billion in compute and training costs snowball out of control. When we take all of these factors, and the expenses hidden by non-GAAP accounting, into account, OpenAI’s losses for 2026 could be substantially larger than 2025. There is a genuine possibility that OpenAI’s 2026 loss will significantly exceed $50 billion (Zitron has even said it could go up to $70 billion), as that is their current trajectory.

This is a slight problem. As Zitron pointed out, OpenAI only has $50 billion in asset value on its books. You might be forgiven for assuming it has more, with all the recent headlines insisting it has raised over $100 billion in investment over the past few months from the likes of Amazon and Nvidia. But the reality is, the vast majority of that investment isn’t cash but effectively vouchers for OpenAI to use these investors’ data centres.

What does all of this mean?

It means that OpenAI could soon rack up losses that exceed its asset value. In other words, they are on course to slam into bankruptcy this year, or at the very latest, early next year, unless they receive tens of billions of dollars of actual cash investment. Even then, that cash injection will likely only delay bankruptcy until late 2027.

This explains why Sam Altman is angling for an IPO as soon as possible, despite internal pushback. But we now know the causes of this internal pushback and why OpenAI is desperate to get government backing. OpenAI needs an IPO to drum up the actual hard cash it requires to delay imminent bankruptcy and keep the ship going for another year. However, with financials this dire, would anyone even buy OpenAI shares? Is an IPO actually possible? Even if OpenAI has enough meme-stock status to allow it to pull the same trick as SpaceX, that will only delay bankruptcy by a year, maybe a year and a half, tops.

These financials are painfully clear — OpenAI is on a trajectory toward imminent collapse. We already knew this. It was obvious to everyone that this scheme was a car crash waiting to happen. This revelation only establishes that the situation is far more dire, and the crunch is coming far sooner, than anyone thought.

---

Thanks for reading! Everything expressed in this article is my opinion, and should not be taken as financial advice or accusations. Don’t forget to check out my YouTube channel for more from me, or Subscribe. Oh, and don’t forget to hit the share button below to get the word out!