--- title: "Forget Spending Caps: One Company Burns Through $500 Million on Claude in a Single Month! -- When AI Becomes Too Expensive to Use" type: "News" locale: "en" url: "https://longbridge.com/en/news/288128600.md" description: "The corporate AI boom is facing a billing crisis, with one company spending $500 million on Claude in a single month due to the lack of spending caps. Tech giants like Microsoft and Amazon are curbing excessive consumption known as \"tokenmaxxing\" by cutting internal AI projects or halting usage tracking. Amazon took offline related tools after employees gamed leaderboards, leading to wasted computing power, emphasizing that AI should not be used for its own sake. The core issue has shifted to cost control and value assessment" datetime: "2026-05-30T01:58:43.000Z" locales: - [zh-CN](https://longbridge.com/zh-CN/news/288128600.md) - [en](https://longbridge.com/en/news/288128600.md) - [zh-HK](https://longbridge.com/zh-HK/news/288128600.md) --- # Forget Spending Caps: One Company Burns Through $500 Million on Claude in a Single Month! -- When AI Becomes Too Expensive to Use The corporate AI boom is encountering its first genuine billing crisis. On May 28, Axios cited an AI consultant who stated that one of their corporate clients recently spent $500 million on Claude in a single month, simply because no limits were set on employee usage. Analysts believe that **many companies, in their rush to roll out AI tools, focused on functionality and promotion while neglecting to establish cost control mechanisms.** Wall Street Insight mentioned that tech giants such as Microsoft and Amazon are taking action to cut internal AI tools or halt projects tracking AI usage, in order to curb excessive consumption behavior known as "tokenmaxxing." An Amazon senior vice president had to issue a warning to employees: > Please do not use AI just for the sake of using AI. The core question facing the market is no longer "whether to embrace AI," but "what exactly have we gained from burning so much money?" **Amazon's case reveals the dilemma of corporate AI governance from another angle.** According to reports citing two insiders, Amazon's developer platform Kiro previously had an internal leaderboard called "Kirorank," which scored employees based on their AI usage activity. However, this leaderboard inadvertently encouraged employees to have AI agents perform meaningless tasks to boost their rankings, directly leading to increased computing power consumption for the company. Amazon Senior Vice President Dave Treadwell admitted to employees this week that while the leaderboard started with good intentions, the end result was that employees drove up the company's operating costs through "tokenmaxxing." He explicitly instructed employees not to focus on token consumption but to concentrate on building better products, emphasizing "do not use AI just for the sake of using AI." **Amazon subsequently confirmed in a statement that the beta dashboard "was not an official or approved tool and has been taken offline."** **Meta faced a similar situation,** where employees also attempted to secure favorable positions in internal rankings by inflating token consumption. This phenomenon indicates that when companies include AI usage in performance assessments, it may backfire, distorting employee incentives into ineffective consumption of computing power. **Amazon has since shifted to using "normalized deployment" metrics instead of token consumption, focusing on tracking whether engineers can continuously generate code with actual value through AI.** Notably, Amazon's capital expenditure is expected to reach $200 billion this year, with the vast majority flowing into AI and data center infrastructure. ## Four Key Issues: Why AI Spending Has Not Delivered Returns According to Axios, corporate AI adoption is facing four structural obstacles. **Misaligned Use Case Selection.** Sophia Velastegui, CEO of Velastegui Ventures and former Chief AI Officer at Microsoft, stated that most people tend to use AI to automate tasks they dislike, rather than those most valuable to the company. She believes that companies should concentrate AI resources on scenarios that directly drive revenue, rather than rolling them out blindly. **Lack of Cost Control.** AI queries are not free; enterprise packages are billed by token. Even simple, routine queries can quickly accumulate into significant expenses, yet most business departments lack a clear understanding of this. **People Are the Biggest Bottleneck.** Velastegui characterized the current widespread "scattergun" approach to AI authorization in enterprises as a path that fails to deliver substantial returns. Companies heap numerous AI tools onto employees but lack effective guidance and focus, resulting in low actual adoption efficiency. **Concerns Over Data Openness.** Josh Pantony, CEO of Boosted.ai, which focuses on AI tools for the financial industry, pointed out that when companies are reluctant to open internal proprietary data to AI agents due to data security concerns, the actual effectiveness of the agents is greatly diminished, making return on investment impossible to achieve. ## Token Economics: The New Core Variable in the AI Narrative Behind this debate lies a more complex investment logic that is being reconstructed. Wall Street Insight mentioned that according to the latest views of Rich Privorotsky, head of Goldman Sachs' One-Delta division, **the core variable in AI trading has shifted from "technical feasibility" to "cost affordability."** DeepSeek reportedly lowered token pricing by 75%, and Xiaomi's MiMo saw price cuts close to 99%. This cost compression could trigger a "price war" logic similar to that seen after subsidy competitions. **He pointed out that infrastructure bottlenecks will eventually ease, and the market should not pay excessive premiums for "problems about to be solved."** Rich Privorotsky further hypothesized whether cheaper tokens would first replace high-cost inference services. If demand expansion experiences a time lag, revenue growth for cloud service providers, model companies, and AI infrastructure may face phased pressure. **He believes that the rationalization of token spending could become a key topic at the board level in the second and third quarters of this year, with importance comparable to the AI growth narrative itself.** According to Bloomberg's Silicon Data LLM Token Expenditure Index, token prices have risen by approximately 65% since late February this year, with cumulative increases in US AI software prices over the past year ranging from 20% to 37%. This cost trend is prompting companies to reevaluate their AI procurement strategies. **As it becomes increasingly feasible to "obtain 90% of the output at 10% of the cost," corporate reliance on high-cost frontier models may systematically decline.** Ali Ansari, CEO of AI model training company Micro1, stated that companies are experiencing a "healthy swing" from overusing AI to using it rationally. He believes: > The only area where AI is currently truly effective is programming. ## Bull-Bear Debate: Same Reality, Two Interpretations Regarding AI investment returns, identical data is pointing to vastly different conclusions under different analytical frameworks. **The bull perspective argues that the current chaos is merely normal growing pains during the transition process.** According to Jim Schneider of Goldman Sachs in early May, by 2030, agentic AI will drive a 24-fold increase in token consumption, and gross margins for hyperscale cloud providers and model providers will turn positive within the next 3 to 12 months. Economic research from JPMorgan also found a surge in Python packages on PyPI in early 2026, a trend that did not appear when ChatGPT launched in 2022, indicating that real productivity improvements are occurring. **The bear perspective was systematically articulated by Goldman Sachs semiconductor analyst Jim Covello in an April report.** He pointed out that almost all value in the AI supply chain is flowing to semiconductor companies, which is historically unprecedented and unsustainable. **Chip companies should benefit when customers benefit, but in this cycle, their prosperity comes at the expense of upstream consumption across the entire industrial chain.** Both narratives are unfolding simultaneously, and the outcome remains unclear. What is certain is that **the simple equation equating "growth in token consumption" with "successful AI transformation" has been broken.** From the extreme case of burning through $500 million in a single month to Amazon halting its gamified leaderboard, AI investments are undergoing stricter scrutiny of returns. How much real value the next AI bill generates will be the true moment of judgment for this high-stakes gamble. Risk Warning and Disclaimer The market carries risks; invest with caution. This article does not constitute personal investment advice, nor does it take into account the specific investment objectives, financial status, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances. Investment decisions made based on this content are the sole responsibility of the investor. ### Related Stocks - [AMZN.US](https://longbridge.com/en/quote/AMZN.US.md) - [MSFT.US](https://longbridge.com/en/quote/MSFT.US.md) - [IXN.US](https://longbridge.com/en/quote/IXN.US.md) - [IDGT.US](https://longbridge.com/en/quote/IDGT.US.md) - [MSFX.US](https://longbridge.com/en/quote/MSFX.US.md) - [DTCR.US](https://longbridge.com/en/quote/DTCR.US.md) - [XDAT.US](https://longbridge.com/en/quote/XDAT.US.md) - [MSFU.US](https://longbridge.com/en/quote/MSFU.US.md) - [CLOU.US](https://longbridge.com/en/quote/CLOU.US.md) - [SRVR.US](https://longbridge.com/en/quote/SRVR.US.md) - [MSFL.US](https://longbridge.com/en/quote/MSFL.US.md) - [AMZZ.US](https://longbridge.com/en/quote/AMZZ.US.md) - [AMZU.US](https://longbridge.com/en/quote/AMZU.US.md) - [AMZW.US](https://longbridge.com/en/quote/AMZW.US.md) - [XSW.US](https://longbridge.com/en/quote/XSW.US.md) - [MSFD.US](https://longbridge.com/en/quote/MSFD.US.md) - [DAT.US](https://longbridge.com/en/quote/DAT.US.md) - [MSFO.US](https://longbridge.com/en/quote/MSFO.US.md) - [AMZD.US](https://longbridge.com/en/quote/AMZD.US.md) - [IGV.US](https://longbridge.com/en/quote/IGV.US.md) - [META.US](https://longbridge.com/en/quote/META.US.md) - [GS.US](https://longbridge.com/en/quote/GS.US.md) - [01810.HK](https://longbridge.com/en/quote/01810.HK.md) - [JPM.US](https://longbridge.com/en/quote/JPM.US.md) - [W4VR.SG](https://longbridge.com/en/quote/W4VR.SG.md) - [XIACY.US](https://longbridge.com/en/quote/XIACY.US.md) - [81810.HK](https://longbridge.com/en/quote/81810.HK.md) - [HXXD.SG](https://longbridge.com/en/quote/HXXD.SG.md) - [JPM-M.US](https://longbridge.com/en/quote/JPM-M.US.md) - [JPM-C.US](https://longbridge.com/en/quote/JPM-C.US.md) - [JPM-D.US](https://longbridge.com/en/quote/JPM-D.US.md) - [JPM-L.US](https://longbridge.com/en/quote/JPM-L.US.md) - [8634.JP](https://longbridge.com/en/quote/8634.JP.md) - [JPM-K.US](https://longbridge.com/en/quote/JPM-K.US.md) - [JPM-J.US](https://longbridge.com/en/quote/JPM-J.US.md) ## Related News & Research - [Was Amazon's Tokenmaxxing Fiasco Behind Claude's $500M Mystery Bill?](https://longbridge.com/en/news/288103609.md) - [Anthropic nears $1 trillion valuation, leapfrogging OpenAI](https://longbridge.com/en/news/287961501.md) - [Anthropic Becomes Most Valuable AI Startup After $65 Bln Funding Round](https://longbridge.com/en/news/288119047.md) - [Update: Market Chatter: Amazon Shuts Internal AI Leaderboard After Employees Inflated Usage Scores](https://longbridge.com/en/news/288107950.md) - [Ex-Snowflake CRO says top engineers don't want this booming AI job](https://longbridge.com/en/news/287801215.md)