Ninety-three percent of enterprises that have moved beyond AI experimentation into active deployment are exceeding their AI budgets, according to a McKinsey survey of 120 enterprise participants published July 20, 2026. The survey found that AI spending increases nearly fourfold as organizations scale from isolated use cases to enterprise-wide adoption, often catching finance and technology leaders off guard.

Sixty-two percent of organizations surveyed have already moved past the experimentation phase into active AI deployment, and a majority expect their AI spend to grow by at least 25 percent over the next 12 months. McKinsey attributed much of the overspending to fragmented purchasing, with business units buying AI capabilities independently and employees building AI-powered workflows outside centralized technology oversight.

Despite the budget pressure, McKinsey found that only 20 to 25 percent of companies have mature AI financial operations practices in place to track and control that spending. Organizations that have implemented active cost optimization measures, such as reusing static prompt context to cut repeated processing costs, have reported savings of 20 to 30 percent on their overall AI spend. About a third of surveyed organizations said they had already achieved savings in that range through specific optimization steps.

McKinsey's researchers said the shift from seat-based software licensing to consumption-based AI pricing is a central driver of the volatility, since usage-based costs expose companies to spending swings that traditional software budgets did not produce.

Source: McKinsey & Company -- https://www.mckinsey.com/capabilities/quantumblack/our-insights/the-cost-of-intelligence-how-cios-can-manage-ai-demand-at-scale