Currently, there are various debates taking place worldwide regarding Artificial Intelligence. However, the hottest topic right now is whether AI is proving to be more expensive than human labour. Millions of layoffs were carried out in the name of AI, yet companies now appear to be in a hard situation. Let’s explore what has transpired over the past few months that has caused tech CEOs to change their tune.
The buzz surrounding AI—or Artificial Intelligence—that has prevailed globally over the past two years was nothing short of a tech revolution. It seemed as though this would be the most significant transformation since the advent of the Internet. Every company was adding “AI” to its name.
According to reports, companies ranging from Google, Microsoft, and Uber to OpenAI have, in one way or another, acknowledged that AI is becoming increasingly expensive. Microsoft has even cancelled several of its Claude licenses because they were proving too costly.
One company received an AI bill amounting to $500 million in just a few months. Similarly, Uber’s AI subscription budget was exhausted ahead of schedule.
A Change in Tune
Every startup was raising capital. Every tech CEO was proclaiming that the world was about to change. But now, in 2026, for the first time, it feels as though the lustre of this narrative has begun to fade slightly.
There isn’t just one reason. Rising expenditure on AI, the lack of a clear revenue model, growing anxieties regarding employment, public backlash, and the shifting narratives from tech companies themselves—all these factors combined are raising a new question: Is the AI bubble slowly beginning to burst?
This question carries significant weight because the tech world has witnessed such a scenario before in history. In the 1990s, a similar atmosphere surrounded the internet. Every company was branding itself as an “internet company.” Investors were pouring in capital without asking many questions. Then, suddenly, the dot-com bubble burst, and thousands of companies vanished.
Costs for Every Query – Has AI Become the Most Expensive Proposition?
Today, a similar sense of unease is palpable regarding AI. The most significant issue is the cost of AI. When ChatGPT first arrived, the general public saw only its magic. People were asking questions, generating images, and having code written for them. However, there was scant discussion regarding the sheer amount of money being burned to power it all.
In reality, AI is no ordinary software; it operates on a massive infrastructure. Behind every query, vast data centers are at work. Thousands of expensive GPU chips consume enormous amounts of electricity, and servers run continuously. This is precisely why the operational cost of running AI continues to rise steadily.
A new term is now gaining rapid currency within the tech industry: “Inference Cost.” Simply put, whenever a user poses a question to an AI or requests the generation of an image, the company incurs a cost in that very instant. The greater the usage, the higher the expenditure.
This is the reason why many companies are now restricting their free AI features. Services that were previously offered for free are now coming with a price tag, and expensive subscription plans are being introduced.
Some reports claim that companies are spending tens of millions of dollars daily just to run video generation models. Several analysts have suggested that the AI industry has become trapped in a phase where its expenditures far outweigh its revenues.
Billions Spent—But Where Are the Returns? Uber COO Andrew Macdonald stated that companies are spending vast sums of money on AI, yet the returns do not appear to be commensurate with the investment. He explicitly noted that, thus far, they have not observed the kind of productivity gains—specifically in terms of work speed and impact—that were originally anticipated.
In other words, for the first time, major corporations are openly asking: Where are the earnings after such massive capital outlay? This question is not being raised solely by small businesses; even the biggest names in the tech world no longer appear as aggressive as they once did.
OpenAI head Sam Altman was among those who predicted that AI would eliminate millions of white-collar jobs. However, his tone now appears to have shifted. He recently acknowledged that the “job apocalypse”—the scenario involving the mass elimination of jobs—has not yet materialised.
In other words, the very people who previously argued that machines would replace humans are now suggesting that humans and AI will work side-by-side.
NVIDIA CEO Jensen Huang also questioned companies that attribute every layoff to AI. He argued that many companies are merely invoking AI to justify their past business decisions.
Google CEO Sundar Pichai also acknowledged that corporate CIOs—the executives responsible for making technology-related decisions—are concerned about the escalating costs associated with AI. He noted that corporate budgets are being depleted rapidly, and this issue could intensify further in the future.
