Capgemini Says 59% Faced Higher AI Costs Than Expected

Capgemini said in an announcement that 59% of executives reported AI spending significantly above expectations or budget during the past year. Among affected organizations, spending exceeded expectations by 24% on average, while 21% went over budget by more than 30%.

Half of executives said AI spending received approval without enough visibility into its full cost. Two thirds of organizations that exceeded their budgets reported financial effects such as lower free cash flow or reduced discretionary spending. In response, 53% tightened budgets, introduced broader spending controls, or reduced the scale or scope of affected projects.

More than three in five executives expect their organizations to adopt a multi model strategy during the next 12 months. The approach would select models based on factors including cost, capability, latency, and risk. Nearly three in five organizations also plan to increase their use of open weight models.

Only one in five organizations calculates AI Yield, which Capgemini defines as verified value after accounting for the full cost of safely producing an AI assisted outcome. The findings come from a September 2026 survey of 500 senior executives at organizations with annual revenue above $1 billion across 10 countries and 13 industries.