Uber's entire annual AI budget was gone by April 2026. This stark reality emerged after Claude Code adoption surged from 32% to 84% across its 5,000-engineer organization between December 2025 and March 2026, according to Cockroachlabs. The rapid embrace of agentic AI tools, intended to boost efficiency, instead created an unforeseen financial drain, depleting funds in a mere three months.
This isn't an isolated incident. Companies are rapidly integrating AI for efficiency and innovation, but operating costs are spiraling out of control. This unchecked adoption creates unpredictable hidden costs, leading to widespread project cancellations and budget crises. Without immediate, sophisticated cost management, many promising AI initiatives will fail, forcing a re-evaluation of AI's true economic value and potentially slowing enterprise adoption.
The Looming Threat of AI Project Cancellations
- More than 40% of agentic AI projects will be canceled by the end of 2027, according to EY. These cancellations stem from escalating costs, unclear business value, or inadequate risk controls. This suggests many organizations are rushing into AI without a clear understanding of its long-term financial and strategic implications, risking significant capital and time.
The Hidden Economics Driving AI's Cost Surge
The brute force, scale-first approach to Large Language Models (LLMs) is reaching an asymptote. Simply adding more data becomes increasingly expensive, making the current economic model for scaling AI unsustainable, according to Ibbaka. Further, current pricing likely understates the long-term economics of agentic AI. Upstream providers may be absorbing, subsidizing, or strategically pricing some compute costs, EY states. This creates a hidden, unsustainable cost structure for adopters, signaling a looming cost cliff when performance limits are hit and subsidies vanish. Companies must look beyond initial pricing to understand the true, long-term financial commitment.
