Amazon Employees Inflate AI Token Consumption
In brief
- Amazon employees have been inflating AI token consumption to meet internal usage targets.
- The company set targets for over 80% of developers to use AI tools each week and tracked consumption on leaderboards.
- Employees used an in-house tool to automate non-essential tasks and maximize token numbers.
- This practice has raised concerns about the reliability of demand figures for AI infrastructure procurement, with combined 2026 capex from top companies tracking between $650 billion and $700 billion.
- The pressure to use AI tools has created perverse incentives, with some employees using the tools to improve their standing on internal leaderboards.
- The future of AI adoption will depend on genuine usage, not inflated numbers.
Terms in this brief
- token consumption
- The amount of computational resources used by AI systems during their operation. This term is crucial for understanding how much infrastructure and processing power companies allocate to AI projects, as excessive or inflated usage can mislead investment decisions in AI technology.
Read full story at Tom's Hardware →, Hacker News →
More briefs
AI Use in Finance Surges
Active use of AI in finance has more than doubled since 2024, rising from 30 percent to 75 percent. Nearly three-quarters of leaders say it is meeting or exceeding ROI expectations. Organizations report significant improvements in decision-making quality, speed, and forecast accuracy. Over 70 percent of organizations report improved decision-making quality and faster decision-making. AI use is expected to continue growing as organizations invest in controls and governance to maintain trust with stakeholders.
AI Boom Creates Huge Demand for Manual Labor
AI is creating a huge demand for manual labor to build its infrastructure. Nvidia CEO says electricians, plumbers, and builders are needed. The investment in AI infrastructure is enormous, with $700 billion expected this year. This could generate almost $7 trillion by the end of the decade. Demand for skilled trades has soared 27% over the past three years. Companies are struggling to hire enough young workers to meet their needs. This labor shortage could lead to six-figure salaries for trades people. Workers will likely benefit from the shortage of skilled labor. New workers will be needed to build AI infrastructure.
AI Profitability Despite High Costs
AI companies are burning through cash at alarming rates, with OpenAI alone spending $25 billion in the first half of 2025 while generating only $4 billion in revenue. The cost of training advanced models continues to rise exponentially due to scaling laws, raising concerns about sustainability. However, despite these high upfront costs, serving existing AI models is highly profitable. For instance, DeepSeek-V4-Pro, a 1.6 trillion-parameter model, costs around $1.74 per million input tokens and $3.48 per million output tokens, yet leading labs charge significantly more-$2-5 per million input tokens and $12-25 per million output tokens. The infrastructure expenses, such as building data centers, are substantial but necessary for future growth. Even if funding dried up, AI would remain a profitable business due to the low marginal cost of serving API calls compared to their pricing. While individual companies face pressure to constantly train new models to stay competitive, the industry as a whole can sustain itself by improving existing models at a steady pace. Looking ahead, the key question is whether current scaling rates can be maintained. If investment dries up, older models will continue to generate profits, potentially allowing incremental improvements to keep up with demand. However, without sufficient funding for new models, competition could erode market share over time.
Reddit Sells Data to Google and OpenAI
Reddit signed a content licensing deal with Google worth around $60 million per year. This deal allows training on Reddit content. An OpenAI partnership followed in 2024. Reddit's revenue mix has three parts: ads, Premium subscriptions, and data licensing. Data licensing is a big part of Reddit's revenue now. It carries high margins because the content already exists. Data licensing revenue will keep growing for Reddit.
Tech Giants Spend $725 Billion on AI
Tech giants are investing a record $725 billion in AI projects. This spending will leave their combined free cash flow down $4 billion in the third quarter. Their free cash flow is on track to fall to the lowest level since 2014. This affects how much cash they have to service debt or return to shareholders. Big Tech companies are making big bets on AI, and it will be interesting to see how this investment pays off in the future.