{
  "id": 20413023,
  "title": "Over half of businesses admit they're facing significantly higher AI costs than they expected",
  "url": "https://urgent.news/2026/11/17/over-half-of-businesses-admit-theyre-facing-significantly-higher-ai",
  "topic": "ai",
  "section": "AI",
  "published": "2026-11-17T09:00:00.000Z",
  "source": {
    "name": "TechRadar",
    "slug": "techradar",
    "url": "https://www.techradar.com/pro/over-half-of-businesses-admit-theyre-facing-significantly-higher-ai-costs-than-they-expected"
  },
  "original_language": "en",
  "account": "Over half of companies are now worried that their expenses for AI technology are much higher than anticipated, according to new research from Capgemini. While the cost of individual tokens used by AI systems is a concern, the research reveals that expenses related to training, governance, and human oversight also play a significant role.\n\nThe survey of 500 CXO-level executives from large organizations found that 53% of those polled are concerned about these AI costs. This concern is mounting even as the price of popular AI models has decreased, as 59% of executives reported that their projected AI-related expenses exceeded their budgets. In fact, 20% of organizations overspent on AI by more than 30%.\n\nAs a result, more businesses are involving their Chief Financial Officers (CFOs) in decisions about AI spending. This shift has been driven by the growing financial burden of AI technology. In response, some companies are implementing multi-model strategies, using various AI models depending on the task at hand.\n\nVolker Darius, CEO of Capgemini Invent, emphasizes that simply controlling costs is not enough to make AI profitable. Organizations must also measure the quality of AI outcomes and the associated business value. Capgemini has introduced the AI Yield metric, which aims to quantify the value of outcomes generated by AI systems. By monitoring AI Yield, companies can make better decisions about allocating capital for AI projects.\n\nThe research suggests that multi-model strategies will become more common in the coming year, with three-quarters of polled executives expecting to adopt them. These strategies will be chosen based on cost, capability, and other factors such as latency and risk. Currently, 60% of organizations measure total AI expenditure against their budgets, but this is expected to decrease to 44% over the next year. Instead, metrics that focus on successful business outcomes from AI use are expected to gain traction.\n\nDarius concludes by stating that as AI becomes more deeply integrated into business workflows, organizations need to treat AI costs as a portfolio of economic outcomes rather than just a technology or procurement issue.",
  "summary": "Organizations are spending more on AI despite a drop in token costs, raising questions on how AI expenditure can be kept under control.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}