Since OpenAI unleashed ChatGPT on the world two and a half years ago, the company has operated at a substantial financial loss.
Despite raising at least $60.9 billion in private funding since ChatGPT’s public launch, OpenAI is leaking billions of dollars every year. In 2024, for example, the tech startup lost some $5 billion, per MSNBC. That doesn’t seem to bother OpenAI insiders, though, who hope to be bringing in $125 billion in annual revenue by 2029.
The gulf between OpenAI’s ambitions and its actual financial health reflect what tech critic Ed Zitron calls the “subprime AI crisis.”
Playing off the subprime mortgage crisis of 2007 — also known as the Great Recession — Zitron notes that AI companies like OpenAI and Anthropic operate at a “massive loss,” and that the chickens are eventually going to come home to roost. Similar to 2007, when banks lent way more credit than they could ever hope to make back, Wall Street has bet billions hoping that OpenAI will miraculously continue to skyrocket in value.
As MSNBC notes, ChatGPT is currently valued at around 30 times its revenue — a staggering amount of faith, to say the least.
Per Zitron, that idea isn’t exactly built on sound financial analysis as much as a frenzy of “magical thinking.”
“I hypothesize a kind of subprime AI crisis is brewing,” Zitron wrote, “where almost the entire tech industry has bought in on a technology sold at a vastly-discounted rate, heavily-centralized and subsidized by big tech. What happens when the entire tech industry relies on the success of a kind of software that only loses money, and doesn’t create much value to begin with?”
The answer to that rhetorical question, as it turns out, is widespread layoffs, price hikes, and the enshittification of software that relies on companies like OpenAI and Anthropic for their own infrastructure.
For example, the company Anysphere, which runs Cursor — an “AI code generator” based on Anthropic’s large language model (LLM) infrastructure — recently imposed a massive rate hike on its users, despite raising nearly $1 billion in funding back in June.
That’s got the company’s dedicated user base of software engineers and coders chuffed, to say the least. Following the hikes, they’ve flocked to social media to vent their frustration, flooding Reddit with posts titled “Cursor: pay more, get less, and don’t ask how it works” and “Cursor’s New Pricing Model Is Absolute Garbage.”
Per Zitron, “Anysphere is, despite getting $900 million in funding, running out of money, or at least believes that continuing to operate its business in a way it did less than a month ago would cause it to do so.”
Following the money upstream reveals the true source of the rate limits: pressure put on Anysphere by Anthropic’s own recent price hikes, the first of their kind, and a dark omen for things to come.
When it comes to OpenAI, it’s difficult to see where else revenue is supposed to come, if not for rate hikes. ChatGPT already boasts 800 million weekly active users, or “something like 10 percent of the world,” according to CEO Sam Altman. The vast majority of those are free users, and it’s difficult to imagine they’d be too happy if their favorite chatbot was suddenly locked behind a paywall.
Case in point, OpenAI has recently introduced similar price hikes to Anthropic’s, offering “priority processing” for businesses willing to pay more.
Regular users have already experienced OpenAI’s random throttling, like in May, when the company limited ChatGPT image requests because their GPUs were “melting” under the weight of its vast user base.
Going forward, the question isn’t a matter of “if” price hikes take off into the stratosphere, but who kicks it off and when.
Zitron sums it up neatly: “There is no way this situation leads to the kind of growth that will make Anthropic and OpenAI profitable, sustainable businesses, and [this] will have the opposite effect on their revenues long-term.”
More on AI: MIT Economist Warns AI Is Poised to Turn Economy Into “Mad Max” Scenario
Share This Article