Several US companies that initially encouraged heavy AI usage are now imposing spending cutbacks and token rationing due to rising costs and unclear value. Early strategies included leaderboards to incentivize maximum AI use, but firms including Amazon and Walmart have since removed such incentives or limited employee token consumption to curb excessive spending [1, 2].

Accenture, which previously linked AI usage to promotion eligibility, has warned employees against wasting AI tokens on trivial tasks like converting PDFs to slides. Justice Kwak of Accenture said, "We’re hitting this inflection point where AI is becoming material to the cost structure. Spend is becoming very unpredictable; and leadership, especially at the CFO, COO, and CIO level, are still asking the question of whether they’re getting value from what we’re spending on in the context of AI" [1].

AI spending has ballooned to billions for many firms, driven by applications such as AI-assisted coding and customer support automation. Uber revealed it spent its entire annual AI budget within the first four months of 2026 and has since introduced monthly spending tiers starting at $1,500, with additional usage requiring prior approval. Uber’s CTO disclosed these limits amid rapid budget exhaustion [3, 2].

Smaller companies are also adjusting. Lindy, a 25-person AI startup, switched from Anthropic Claude models to a cheaper Chinese provider DeepSeek to drastically reduce costs. CEO Flo Crivello called it "a matter of survival for the business. That's all it is" [3]. Meanwhile, McKinsey disclosed using approximately 25,000 AI agents alongside 40,000 human employees, aiming to support every employee with AI tools [2].

Some executives question the sustainability of the current AI business model. Cisco's Jeetu Patel said, "the price is far higher than the actual value these tokens are generating at scale" [1, 2]. Russell Fradin, CEO of Larridin, noted that while companies are not necessarily planning to spend less on AI next year, they are focusing on "instrumenting to understand where it goes" [2].

Amazon removed its AI token usage leaderboard to prevent overspending, and Walmart imposed token limits on its internally developed AI tools [2].

In June 2026, Uber’s spending caps and tiered approach came into effect following the rapid burnout of their initial budget [3, 2]. Startups like Lindy also completed AI vendor switches this month to control costs [3]. McKinsey's AI deployment data were released earlier in 2026 [2].