NewsUber's AI Budget Crisis Reveals 'Tokenmaxxing' - AI Cost Without ROI at Scale
VentureBeat reported on what it termed "tokenmaxxing" - a pattern emerging at Uber and other large technology organizations where AI token usage is surging without a corresponding increase in business outcomes. Enterprises are consuming more AI compute than ever while struggling to optimize the cost-to-return ratio of that consumption. Uber's AI budget crisis illustrated the pattern at scale: increasing AI spend, flat measurable returns, and growing pressure to justify the investment to finance leadership.
The pattern is consistent with the broader ROI data from Deloitte, McKinsey, and Domino: AI is generating activity - tokens consumed, queries processed, outputs generated - without generating proportional business value. The gap between AI activity and AI outcomes is the defining challenge of enterprise AI in 2026.
Why it matters
For organizations evaluating AI investments, the tokenmaxxing pattern is a warning signal. AI spend that is measured in activity - usage, adoption, prompts - rather than in outcomes is almost always disconnected from the business problem it is supposed to solve. The organizations escaping this trap are the ones that defined the outcome before they defined the technology.
