# Uber Overcomes AI Budget Overrun with New Strategies The company declares the end of the tokenmaxxing era. - **Source:** [Fortune](https://fortune.com) - **Published:** 2026-08-07 - **Topic:** [Data & AI](https://brifing.si/c/data-ai) - **Original article:** - **On Brifing:** https://brifing.si/s/40a7de1b-2ccd-428e-8aa8-4d48162bc3e3 ## Uber's AI Budget Depleted Uber exhausted its AI budget for 2026 within the first few months of the year. The company encouraged employees to use tools like Anthropic’s Claude Code extensively, even creating leaderboards to rank software engineers based on usage. This highlighted the need for better budget control in AI spending. ## End of the Tokenmaxxing Era Uber CTO Praveen Neppalli Naga announced the end of the tokenmaxxing era, where companies incentivized AI usage. Uber optimized efficiency by improving prompt caching and adjusting default model settings, reducing costs significantly. ## AI Costs Are Decreasing Uber quadrupled the number of employees using frontier AI tools, lowering the cost per token. Engineers were given visibility into their AI usage and hourly costs, leading to increased efficiency and reduced overall spending. ## ROI on AI Investments According to Deutsche Bank Research Institute, ROI on AI investments is not yet at expected levels. While Magnificent Seven companies saw profit margins rise from 15% to 25% between 2023-2026, the rest of the S&P 500 saw only a 10% increase, indicating limited widespread returns. ## Risk of Jevons Paradox Uber faces the risk of Jevons paradox as it reduces token costs. This paradox, observed in 19th-century coal consumption, suggests that cheaper resources lead to increased spending. The same dynamic could apply to AI tokens. ## Rising Token Expenditure The Silicon Data Token Expenditure Index shows that while token prices dropped over 90% since 2023, spending on large language models doubled since late last year. Companies are using more AI tools as costs decrease. ## Focus on Quality Uber adopted a philosophy prioritizing quality over quantity in token spending. By making AI usage more efficient, the company managed to control costs effectively, setting an example for future AI applications. ## The Future of AI Uber’s CTO emphasized that the future of AI will be shaped by companies that use tokens most efficiently, not those that spend the most. This new phase requires smarter and more sustainable management of AI investments. ## Sources covering this story - https://fortune.com/2026/08/07/uber-ai-spending-tokenmaxxing-is-over-cto/ — --- _Summary written by Brifing from the original reporting. Copyright in the original article belongs to its publisher; read the full piece at the link above._