In November 2025, I published an analysis on Channel 10 explaining why an OpenAI collapse is a perfectly plausible scenario. In the time since, it seems like it would take a genuine miracle for that scenario not to materialize, and the shockwaves from such an event could even trigger a global recession.
-
Around that time, Sam Altman was strutting around Asia like a peacock, scribbling his signature on every proposal put in front of him, regardless of the price tag attached. He returned to the United States as a conqueror, his briefcase stuffed with commitments detached from reality worth more than a trillion dollars — all while the company posted a net annual loss of nearly $40 billion.
A net loss for a growing company with hundreds of millions of users isn't a problem in itself. Amazon, Netflix, and Tesla operated this way for years, and they're still excellent companies. The problem is that even then, it was clear OpenAI was in trouble that would be hard to solve.
-
For an AI company to operate, four components are required: processing hardware for training and running models, quality data pools to train those models on, cloud infrastructure to serve users, and an existing base of products and users that, on one hand, generates independent revenue, and on the other, enables rapid distribution of AI capabilities across the user base of the company's existing products.
Unfortunately, OpenAI doesn't have even one of these basic requirements. This means the loss isn't from infrastructure investment — because there is no infrastructure. The loss comes from delivering the product itself: the more users the product has, the deeper the loss gets. The company promises to finance the debt through growth, but if the price of that growth is a deepening of the debt, then this growth will only hasten the end rather than postpone it. On top of all this, the company made a strategic mistake by choosing to focus on the consumer market. The individual consumer is price-sensitive and churns at the click of a button, unlike Anthropic's enterprise customers.
-
Sam Altman looks at the history of his nemesis, Elon Musk, and thinks they're the same. Elon always knew how to convince investors to hand over more and more money, and things somehow worked out in the end. What Sam doesn't understand is the secret of Musk's magic trick — mastery of addition and subtraction in their simplest sense.
Musk's companies have sound economic logic, and although they face unbelievable challenges, meeting those challenges promises massive profits. In OpenAI's case, meeting its targets promises insolvency and collapse — and that makes all the difference.
-
An OpenAI collapse would wipe out several small companies that depend on it as a primary customer. Among these are Cerebras (which might survive thanks to a successful IPO) and Coreweave. Other giant companies are also at enormous risk due to their exposure through large-scale contracts with OpenAI, including Oracle, Broadcom, SoftBank, Microsoft, and of course Nvidia.
Given that the growth of the American economy in 2025 was based almost entirely on the AI industry, such a scenario could lead to a recession of some magnitude, at least until supply and demand reorganize around the surviving providers.
The last word hasn't been said, and it's possible the company's major investors will manage to keep it on life support long enough to prevent even greater losses were it to collapse. But the delay of the company's IPO is no accident — it's clear that a failed IPO would seal both its fate and the fate of the entire economy.
No one can predict the future, but despite the sudden attractiveness of chip stocks on Wall Street, this might be a blazing warning sign rather than an opportunity.
In the image: a server farm (Google's. OpenAI doesn't have one...)
--
👋 Hi, I'm Shlomo Strauss, and my posts are not written by artificial intelligence.
Feel free to follow me for more interesting content on science and technology, or search for "Shlomo Strauss's Blog" on Google.