Uber has faced a significant challenge with its artificial intelligence investments. The company rapidly exhausted its annual AI budget within four months, raising questions about the effectiveness of these expenditures. In December 2025, Uber deployed Anthropic’s Claude Code to around 5,000 engineers. However, by April 2026, the allocated funds for AI tools, including Claude Code and Cursor, were completely depleted.
The statistics regarding AI engagement among engineers at Uber are impressive, yet the actual returns appear uncertain. By spring 2026, 95% of engineers utilized AI tools monthly, and roughly 70% of code commits were AI-based. Significantly, the use of AI-driven features soared from 32% in February to 84% by March. However, the company’s COO admitted that connecting these figures to tangible outcomes remains elusive, as it is challenging to link the usage of AI to increased production of consumer features.
Uber’s escalating costs shed light on the broader issues enterprises face regarding budgeting for AI tools. Monthly API expenses for engineers varied widely, from $500 to $2,000 each. Uber's research and development spending reached $3.4 billion in 2025, up 9% from the previous year. Initially intended to sustain through 2026, the AI budget lasted a mere four months.
The realization of these financial outcomes prompted Uber's CTO to disclose internal reviews concerning AI investments. The company now grapples with difficult decisions about whether to continue increasing investments in artificial intelligence or reassess its workforce in response to skyrocketing costs. Alarmingly, 11% of backend updates were performed by AI without any human oversight.
Uber is not alone in facing this predicament, as others in the tech industry, such as Microsoft, have also restricted the use of AI due to rising expenses. For investors observing the intersection of AI and cryptocurrency, this situation reveals significant insights about variable token pricing and its implications for enterprise budgets. Many crypto-based projects, such as Akash and Render, position themselves as cost-effective alternatives to traditional AI infrastructure, highlighting a dynamic landscape in demand for computational resources in the AI sector.
Uber's experience serves as a cautionary tale, demonstrating that while demand for AI solutions may increase rapidly, the financial implications can result in budget limitations much sooner than anticipated. Investors must remain aware that aspiring demands might not correlate with sustainable growth under current economic parameters.