The artificial intelligence revolution aims to bring unprecedented efficiency and unlimited productivity to the technological landscape by cutting costs.
But the promise of revolution is currently hitting a harsh financial reality marked by unexpected AI-related costs, ruining the budget tech companies designed for the year 2026.
Of all the companies, Uber and Microsoft are struggling with AI budget overruns. For example, Microsoft recently decided to revoke its internal cloud code licenses as well as end access to its Experiences and Devices division by June 30, 2026.
As reported by Fortune, the decision to cancel the pilot has been taken just 6 months after its launch in December 2025 due to unstable token-based billing consumption.
The announcement comes after AI costs exceeded company expectations. Engineers are advised to use GitHub Copilot CLI instead of Cloud.
Similarly, Uber has decided to reduce AI technologies due to increased budget due to excessive consumption.
Uber’s chief technology officer, Parveen Nappalli Naga, disclosed the crisis as the company exhausted its $3.4 billion 2026 budget by April.
The company introduced cloud code to 5000 engineers. Within four months, usage increased to unprecedented levels. Monthly usage rates among engineers increased 85–95 percent and API costs were between $500 and $2000 monthly per engineer.
Similarly, Nvidia’s vice president of applied deep learning also acknowledged that the cost of AI has exceeded human-based costs.
talking to axios“For my team, the cost of computation far outweighs the cost of staff,” said Brian Catanzaro.
The current AI cost crisis stems from how these coding agents are being deployed in tech companies. Of all the factors responsible for the cost explosion, token-based scaling is the most prominent as thousands of developers use AI systems to automate large amounts of their coding.
Given the intensity of the crisis, the debate over the bursting of the AI bubble has once again erupted. According to experts, we are moving into the reality check phase.
Tech-savvy users also shared their opinions on the social media platform, with one writing, “It’s an AI bubble- ok AI can help solve your problems partially but humans are very much needed at every stage of the life cycle.”
Another commented, “The cost problem is real, and yet no one is talking loudly enough about it. Everyone has tracking capabilities, but almost no one is keeping track of the bills.”
As maintenance costs continue to rise, the question is no longer just whether AI will change the world, but also whether the current business models powering it are standing on a bubble ready to burst.
